representations definition in science

• Open access
• Published: 19 July 2015

The role of visual representations in scientific practices: from conceptual understanding and knowledge generation to ‘seeing’ how science works

• Maria Evagorou 1 ,
• Sibel Erduran 2 &
• Terhi Mäntylä 3  

International Journal of STEM Education volume  2 , Article number:  11 ( 2015 ) Cite this article

72k Accesses

78 Citations

16 Altmetric

Metrics details

The use of visual representations (i.e., photographs, diagrams, models) has been part of science, and their use makes it possible for scientists to interact with and represent complex phenomena, not observable in other ways. Despite a wealth of research in science education on visual representations, the emphasis of such research has mainly been on the conceptual understanding when using visual representations and less on visual representations as epistemic objects. In this paper, we argue that by positioning visual representations as epistemic objects of scientific practices, science education can bring a renewed focus on how visualization contributes to knowledge formation in science from the learners’ perspective.

This is a theoretical paper, and in order to argue about the role of visualization, we first present a case study, that of the discovery of the structure of DNA that highlights the epistemic components of visual information in science. The second case study focuses on Faraday’s use of the lines of magnetic force. Faraday is known of his exploratory, creative, and yet systemic way of experimenting, and the visual reasoning leading to theoretical development was an inherent part of the experimentation. Third, we trace a contemporary account from science focusing on the experimental practices and how reproducibility of experimental procedures can be reinforced through video data.

Conclusions

Our conclusions suggest that in teaching science, the emphasis in visualization should shift from cognitive understanding—using the products of science to understand the content—to engaging in the processes of visualization. Furthermore, we suggest that is it essential to design curriculum materials and learning environments that create a social and epistemic context and invite students to engage in the practice of visualization as evidence, reasoning, experimental procedure, or a means of communication and reflect on these practices. Implications for teacher education include the need for teacher professional development programs to problematize the use of visual representations as epistemic objects that are part of scientific practices.

During the last decades, research and reform documents in science education across the world have been calling for an emphasis not only on the content but also on the processes of science (Bybee 2014 ; Eurydice 2012 ; Duschl and Bybee 2014 ; Osborne 2014 ; Schwartz et al. 2012 ), in order to make science accessible to the students and enable them to understand the epistemic foundation of science. Scientific practices, part of the process of science, are the cognitive and discursive activities that are targeted in science education to develop epistemic understanding and appreciation of the nature of science (Duschl et al. 2008 ) and have been the emphasis of recent reform documents in science education across the world (Achieve 2013 ; Eurydice 2012 ). With the term scientific practices, we refer to the processes that take place during scientific discoveries and include among others: asking questions, developing and using models, engaging in arguments, and constructing and communicating explanations (National Research Council 2012 ). The emphasis on scientific practices aims to move the teaching of science from knowledge to the understanding of the processes and the epistemic aspects of science. Additionally, by placing an emphasis on engaging students in scientific practices, we aim to help students acquire scientific knowledge in meaningful contexts that resemble the reality of scientific discoveries.

Despite a wealth of research in science education on visual representations, the emphasis of such research has mainly been on the conceptual understanding when using visual representations and less on visual representations as epistemic objects. In this paper, we argue that by positioning visual representations as epistemic objects, science education can bring a renewed focus on how visualization contributes to knowledge formation in science from the learners’ perspective. Specifically, the use of visual representations (i.e., photographs, diagrams, tables, charts) has been part of science and over the years has evolved with the new technologies (i.e., from drawings to advanced digital images and three dimensional models). Visualization makes it possible for scientists to interact with complex phenomena (Richards 2003 ), and they might convey important evidence not observable in other ways. Visual representations as a tool to support cognitive understanding in science have been studied extensively (i.e., Gilbert 2010 ; Wu and Shah 2004 ). Studies in science education have explored the use of images in science textbooks (i.e., Dimopoulos et al. 2003 ; Bungum 2008 ), students’ representations or models when doing science (i.e., Gilbert et al. 2008 ; Dori et al. 2003 ; Lehrer and Schauble 2012 ; Schwarz et al. 2009 ), and students’ images of science and scientists (i.e., Chambers 1983 ). Therefore, studies in the field of science education have been using the term visualization as “the formation of an internal representation from an external representation” (Gilbert et al. 2008 , p. 4) or as a tool for conceptual understanding for students.

In this paper, we do not refer to visualization as mental image, model, or presentation only (Gilbert et al. 2008 ; Philips et al. 2010 ) but instead focus on visual representations or visualization as epistemic objects. Specifically, we refer to visualization as a process for knowledge production and growth in science. In this respect, modeling is an aspect of visualization, but what we are focusing on with visualization is not on the use of model as a tool for cognitive understanding (Gilbert 2010 ; Wu and Shah 2004 ) but the on the process of modeling as a scientific practice which includes the construction and use of models, the use of other representations, the communication in the groups with the use of the visual representation, and the appreciation of the difficulties that the science phase in this process. Therefore, the purpose of this paper is to present through the history of science how visualization can be considered not only as a cognitive tool in science education but also as an epistemic object that can potentially support students to understand aspects of the nature of science.

Scientific practices and science education

According to the New Generation Science Standards (Achieve 2013 ), scientific practices refer to: asking questions and defining problems; developing and using models; planning and carrying out investigations; analyzing and interpreting data; using mathematical and computational thinking; constructing explanations and designing solutions; engaging in argument from evidence; and obtaining, evaluating, and communicating information. A significant aspect of scientific practices is that science learning is more than just about learning facts, concepts, theories, and laws. A fuller appreciation of science necessitates the understanding of the science relative to its epistemological grounding and the process that are involved in the production of knowledge (Hogan and Maglienti 2001 ; Wickman 2004 ).

The New Generation Science Standards is, among other changes, shifting away from science inquiry and towards the inclusion of scientific practices (Duschl and Bybee 2014 ; Osborne 2014 ). By comparing the abilities to do scientific inquiry (National Research Council 2000 ) with the set of scientific practices, it is evident that the latter is about engaging in the processes of doing science and experiencing in that way science in a more authentic way. Engaging in scientific practices according to Osborne ( 2014 ) “presents a more authentic picture of the endeavor that is science” (p.183) and also helps the students to develop a deeper understanding of the epistemic aspects of science. Furthermore, as Bybee ( 2014 ) argues, by engaging students in scientific practices, we involve them in an understanding of the nature of science and an understanding on the nature of scientific knowledge.

Science as a practice and scientific practices as a term emerged by the philosopher of science, Kuhn (Osborne 2014 ), refers to the processes in which the scientists engage during knowledge production and communication. The work that is followed by historians, philosophers, and sociologists of science (Latour 2011 ; Longino 2002 ; Nersessian 2008 ) revealed the scientific practices in which the scientists engage in and include among others theory development and specific ways of talking, modeling, and communicating the outcomes of science.

Visualization as an epistemic object

Schematic, pictorial symbols in the design of scientific instruments and analysis of the perceptual and functional information that is being stored in those images have been areas of investigation in philosophy of scientific experimentation (Gooding et al. 1993 ). The nature of visual perception, the relationship between thought and vision, and the role of reproducibility as a norm for experimental research form a central aspect of this domain of research in philosophy of science. For instance, Rothbart ( 1997 ) has argued that visualizations are commonplace in the theoretical sciences even if every scientific theory may not be defined by visualized models.

Visual representations (i.e., photographs, diagrams, tables, charts, models) have been used in science over the years to enable scientists to interact with complex phenomena (Richards 2003 ) and might convey important evidence not observable in other ways (Barber et al. 2006 ). Some authors (e.g., Ruivenkamp and Rip 2010 ) have argued that visualization is as a core activity of some scientific communities of practice (e.g., nanotechnology) while others (e.g., Lynch and Edgerton 1988 ) have differentiated the role of particular visualization techniques (e.g., of digital image processing in astronomy). Visualization in science includes the complex process through which scientists develop or produce imagery, schemes, and graphical representation, and therefore, what is of importance in this process is not only the result but also the methodology employed by the scientists, namely, how this result was produced. Visual representations in science may refer to objects that are believed to have some kind of material or physical existence but equally might refer to purely mental, conceptual, and abstract constructs (Pauwels 2006 ). More specifically, visual representations can be found for: (a) phenomena that are not observable with the eye (i.e., microscopic or macroscopic); (b) phenomena that do not exist as visual representations but can be translated as such (i.e., sound); and (c) in experimental settings to provide visual data representations (i.e., graphs presenting velocity of moving objects). Additionally, since science is not only about replicating reality but also about making it more understandable to people (either to the public or other scientists), visual representations are not only about reproducing the nature but also about: (a) functioning in helping solving a problem, (b) filling gaps in our knowledge, and (c) facilitating knowledge building or transfer (Lynch 2006 ).

Using or developing visual representations in the scientific practice can range from a straightforward to a complicated situation. More specifically, scientists can observe a phenomenon (i.e., mitosis) and represent it visually using a picture or diagram, which is quite straightforward. But they can also use a variety of complicated techniques (i.e., crystallography in the case of DNA studies) that are either available or need to be developed or refined in order to acquire the visual information that can be used in the process of theory development (i.e., Latour and Woolgar 1979 ). Furthermore, some visual representations need decoding, and the scientists need to learn how to read these images (i.e., radiologists); therefore, using visual representations in the process of science requires learning a new language that is specific to the medium/methods that is used (i.e., understanding an X-ray picture is different from understanding an MRI scan) and then communicating that language to other scientists and the public.

There are much intent and purposes of visual representations in scientific practices, as for example to make a diagnosis, compare, describe, and preserve for future study, verify and explore new territory, generate new data (Pauwels 2006 ), or present new methodologies. According to Latour and Woolgar ( 1979 ) and Knorr Cetina ( 1999 ), visual representations can be used either as primary data (i.e., image from a microscope). or can be used to help in concept development (i.e., models of DNA used by Watson and Crick), to uncover relationships and to make the abstract more concrete (graphs of sound waves). Therefore, visual representations and visual practices, in all forms, are an important aspect of the scientific practices in developing, clarifying, and transmitting scientific knowledge (Pauwels 2006 ).

Methods and Results: Merging Visualization and scientific practices in science

In this paper, we present three case studies that embody the working practices of scientists in an effort to present visualization as a scientific practice and present our argument about how visualization is a complex process that could include among others modeling and use of representation but is not only limited to that. The first case study explores the role of visualization in the construction of knowledge about the structure of DNA, using visuals as evidence. The second case study focuses on Faraday’s use of the lines of magnetic force and the visual reasoning leading to the theoretical development that was an inherent part of the experimentation. The third case study focuses on the current practices of scientists in the context of a peer-reviewed journal called the Journal of Visualized Experiments where the methodology is communicated through videotaped procedures. The three case studies represent the research interests of the three authors of this paper and were chosen to present how visualization as a practice can be involved in all stages of doing science, from hypothesizing and evaluating evidence (case study 1) to experimenting and reasoning (case study 2) to communicating the findings and methodology with the research community (case study 3), and represent in this way the three functions of visualization as presented by Lynch ( 2006 ). Furthermore, the last case study showcases how the development of visualization technologies has contributed to the communication of findings and methodologies in science and present in that way an aspect of current scientific practices. In all three cases, our approach is guided by the observation that the visual information is an integral part of scientific practices at the least and furthermore that they are particularly central in the scientific practices of science.

Case study 1: use visual representations as evidence in the discovery of DNA

The focus of the first case study is the discovery of the structure of DNA. The DNA was first isolated in 1869 by Friedrich Miescher, and by the late 1940s, it was known that it contained phosphate, sugar, and four nitrogen-containing chemical bases. However, no one had figured the structure of the DNA until Watson and Crick presented their model of DNA in 1953. Other than the social aspects of the discovery of the DNA, another important aspect was the role of visual evidence that led to knowledge development in the area. More specifically, by studying the personal accounts of Watson ( 1968 ) and Crick ( 1988 ) about the discovery of the structure of the DNA, the following main ideas regarding the role of visual representations in the production of knowledge can be identified: (a) The use of visual representations was an important part of knowledge growth and was often dependent upon the discovery of new technologies (i.e., better microscopes or better techniques in crystallography that would provide better visual representations as evidence of the helical structure of the DNA); and (b) Models (three-dimensional) were used as a way to represent the visual images (X-ray images) and connect them to the evidence provided by other sources to see whether the theory can be supported. Therefore, the model of DNA was built based on the combination of visual evidence and experimental data.

An example showcasing the importance of visual representations in the process of knowledge production in this case is provided by Watson, in his book The Double Helix (1968):

…since the middle of the summer Rosy [Rosalind Franklin] had had evidence for a new three-dimensional form of DNA. It occurred when the DNA 2molecules were surrounded by a large amount of water. When I asked what the pattern was like, Maurice went into the adjacent room to pick up a print of the new form they called the “B” structure. The instant I saw the picture, my mouth fell open and my pulse began to race. The pattern was unbelievably simpler than those previously obtained (A form). Moreover, the black cross of reflections which dominated the picture could arise only from a helical structure. With the A form the argument for the helix was never straightforward, and considerable ambiguity existed as to exactly which type of helical symmetry was present. With the B form however, mere inspection of its X-ray picture gave several of the vital helical parameters. (p. 167-169)

As suggested by Watson’s personal account of the discovery of the DNA, the photo taken by Rosalind Franklin (Fig.  1 ) convinced him that the DNA molecule must consist of two chains arranged in a paired helix, which resembles a spiral staircase or ladder, and on March 7, 1953, Watson and Crick finished and presented their model of the structure of DNA (Watson and Berry 2004 ; Watson 1968 ) which was based on the visual information provided by the X-ray image and their knowledge of chemistry.

X-ray chrystallography of DNA

In analyzing the visualization practice in this case study, we observe the following instances that highlight how the visual information played a role:

Asking questions and defining problems: The real world in the model of science can at some points only be observed through visual representations or representations, i.e., if we are using DNA as an example, the structure of DNA was only observable through the crystallography images produced by Rosalind Franklin in the laboratory. There was no other way to observe the structure of DNA, therefore the real world.

Analyzing and interpreting data: The images that resulted from crystallography as well as their interpretations served as the data for the scientists studying the structure of DNA.

Experimenting: The data in the form of visual information were used to predict the possible structure of the DNA.

Modeling: Based on the prediction, an actual three-dimensional model was prepared by Watson and Crick. The first model did not fit with the real world (refuted by Rosalind Franklin and her research group from King’s College) and Watson and Crick had to go through the same process again to find better visual evidence (better crystallography images) and create an improved visual model.

Example excerpts from Watson’s biography provide further evidence for how visualization practices were applied in the context of the discovery of DNA (Table  1 ).

In summary, by examining the history of the discovery of DNA, we showcased how visual data is used as scientific evidence in science, identifying in that way an aspect of the nature of science that is still unexplored in the history of science and an aspect that has been ignored in the teaching of science. Visual representations are used in many ways: as images, as models, as evidence to support or rebut a model, and as interpretations of reality.

Case study 2: applying visual reasoning in knowledge production, the example of the lines of magnetic force

The focus of this case study is on Faraday’s use of the lines of magnetic force. Faraday is known of his exploratory, creative, and yet systemic way of experimenting, and the visual reasoning leading to theoretical development was an inherent part of this experimentation (Gooding 2006 ). Faraday’s articles or notebooks do not include mathematical formulations; instead, they include images and illustrations from experimental devices and setups to the recapping of his theoretical ideas (Nersessian 2008 ). According to Gooding ( 2006 ), “Faraday’s visual method was designed not to copy apparent features of the world, but to analyse and replicate them” (2006, p. 46).

The lines of force played a central role in Faraday’s research on electricity and magnetism and in the development of his “field theory” (Faraday 1852a ; Nersessian 1984 ). Before Faraday, the experiments with iron filings around magnets were known and the term “magnetic curves” was used for the iron filing patterns and also for the geometrical constructs derived from the mathematical theory of magnetism (Gooding et al. 1993 ). However, Faraday used the lines of force for explaining his experimental observations and in constructing the theory of forces in magnetism and electricity. Examples of Faraday’s different illustrations of lines of magnetic force are given in Fig.  2 . Faraday gave the following experiment-based definition for the lines of magnetic forces:

a Iron filing pattern in case of bar magnet drawn by Faraday (Faraday 1852b , Plate IX, p. 158, Fig. 1), b Faraday’s drawing of lines of magnetic force in case of cylinder magnet, where the experimental procedure, knife blade showing the direction of lines, is combined into drawing (Faraday, 1855, vol. 1, plate 1)

A line of magnetic force may be defined as that line which is described by a very small magnetic needle, when it is so moved in either direction correspondent to its length, that the needle is constantly a tangent to the line of motion; or it is that line along which, if a transverse wire be moved in either direction, there is no tendency to the formation of any current in the wire, whilst if moved in any other direction there is such a tendency; or it is that line which coincides with the direction of the magnecrystallic axis of a crystal of bismuth, which is carried in either direction along it. The direction of these lines about and amongst magnets and electric currents, is easily represented and understood, in a general manner, by the ordinary use of iron filings. (Faraday 1852a , p. 25 (3071))

The definition describes the connection between the experiments and the visual representation of the results. Initially, the lines of force were just geometric representations, but later, Faraday treated them as physical objects (Nersessian 1984 ; Pocovi and Finlay 2002 ):

I have sometimes used the term lines of force so vaguely, as to leave the reader doubtful whether I intended it as a merely representative idea of the forces, or as the description of the path along which the power was continuously exerted. … wherever the expression line of force is taken simply to represent the disposition of forces, it shall have the fullness of that meaning; but that wherever it may seem to represent the idea of the physical mode of transmission of the force, it expresses in that respect the opinion to which I incline at present. The opinion may be erroneous, and yet all that relates or refers to the disposition of the force will remain the same. (Faraday, 1852a , p. 55-56 (3075))

He also felt that the lines of force had greater explanatory power than the dominant theory of action-at-a-distance:

Now it appears to me that these lines may be employed with great advantage to represent nature, condition, direction and comparative amount of the magnetic forces; and that in many cases they have, to the physical reasoned at least, a superiority over that method which represents the forces as concentrated in centres of action… (Faraday, 1852a , p. 26 (3074))

For giving some insight to Faraday’s visual reasoning as an epistemic practice, the following examples of Faraday’s studies of the lines of magnetic force (Faraday 1852a , 1852b ) are presented:

(a) Asking questions and defining problems: The iron filing patterns formed the empirical basis for the visual model: 2D visualization of lines of magnetic force as presented in Fig.  2 . According to Faraday, these iron filing patterns were suitable for illustrating the direction and form of the magnetic lines of force (emphasis added):

It must be well understood that these forms give no indication by their appearance of the relative strength of the magnetic force at different places, inasmuch as the appearance of the lines depends greatly upon the quantity of filings and the amount of tapping; but the direction and forms of these lines are well given, and these indicate, in a considerable degree, the direction in which the forces increase and diminish . (Faraday 1852b , p.158 (3237))

Despite being static and two dimensional on paper, the lines of magnetic force were dynamical (Nersessian 1992 , 2008 ) and three dimensional for Faraday (see Fig.  2 b). For instance, Faraday described the lines of force “expanding”, “bending,” and “being cut” (Nersessian 1992 ). In Fig.  2 b, Faraday has summarized his experiment (bar magnet and knife blade) and its results (lines of force) in one picture.

(b) Analyzing and interpreting data: The model was so powerful for Faraday that he ended up thinking them as physical objects (e.g., Nersessian 1984 ), i.e., making interpretations of the way forces act. Of course, he made a lot of experiments for showing the physical existence of the lines of force, but he did not succeed in it (Nersessian 1984 ). The following quote illuminates Faraday’s use of the lines of force in different situations:

The study of these lines has, at different times, been greatly influential in leading me to various results, which I think prove their utility as well as fertility. Thus, the law of magneto-electric induction; the earth’s inductive action; the relation of magnetism and light; diamagnetic action and its law, and magnetocrystallic action, are the cases of this kind… (Faraday 1852a , p. 55 (3174))

(c) Experimenting: In Faraday's case, he used a lot of exploratory experiments; in case of lines of magnetic force, he used, e.g., iron filings, magnetic needles, or current carrying wires (see the quote above). The magnetic field is not directly observable and the representation of lines of force was a visual model, which includes the direction, form, and magnitude of field.

(d) Modeling: There is no denying that the lines of magnetic force are visual by nature. Faraday’s views of lines of force developed gradually during the years, and he applied and developed them in different contexts such as electromagnetic, electrostatic, and magnetic induction (Nersessian 1984 ). An example of Faraday’s explanation of the effect of the wire b’s position to experiment is given in Fig.  3 . In Fig.  3 , few magnetic lines of force are drawn, and in the quote below, Faraday is explaining the effect using these magnetic lines of force (emphasis added):

Picture of an experiment with different arrangements of wires ( a , b’ , b” ), magnet, and galvanometer. Note the lines of force drawn around the magnet. (Faraday 1852a , p. 34)

It will be evident by inspection of Fig. 3 , that, however the wires are carried away, the general result will, according to the assumed principles of action, be the same; for if a be the axial wire, and b’, b”, b”’ the equatorial wire, represented in three different positions, whatever magnetic lines of force pass across the latter wire in one position, will also pass it in the other, or in any other position which can be given to it. The distance of the wire at the place of intersection with the lines of force, has been shown, by the experiments (3093.), to be unimportant. (Faraday 1852a , p. 34 (3099))

In summary, by examining the history of Faraday’s use of lines of force, we showed how visual imagery and reasoning played an important part in Faraday’s construction and representation of his “field theory”. As Gooding has stated, “many of Faraday’s sketches are far more that depictions of observation, they are tools for reasoning with and about phenomena” (2006, p. 59).

Case study 3: visualizing scientific methods, the case of a journal

The focus of the third case study is the Journal of Visualized Experiments (JoVE) , a peer-reviewed publication indexed in PubMed. The journal devoted to the publication of biological, medical, chemical, and physical research in a video format. The journal describes its history as follows:

JoVE was established as a new tool in life science publication and communication, with participation of scientists from leading research institutions. JoVE takes advantage of video technology to capture and transmit the multiple facets and intricacies of life science research. Visualization greatly facilitates the understanding and efficient reproduction of both basic and complex experimental techniques, thereby addressing two of the biggest challenges faced by today's life science research community: i) low transparency and poor reproducibility of biological experiments and ii) time and labor-intensive nature of learning new experimental techniques. ( http://www.jove.com/ )

By examining the journal content, we generate a set of categories that can be considered as indicators of relevance and significance in terms of epistemic practices of science that have relevance for science education. For example, the quote above illustrates how scientists view some norms of scientific practice including the norms of “transparency” and “reproducibility” of experimental methods and results, and how the visual format of the journal facilitates the implementation of these norms. “Reproducibility” can be considered as an epistemic criterion that sits at the heart of what counts as an experimental procedure in science:

Investigating what should be reproducible and by whom leads to different types of experimental reproducibility, which can be observed to play different roles in experimental practice. A successful application of the strategy of reproducing an experiment is an achievement that may depend on certain isiosyncratic aspects of a local situation. Yet a purely local experiment that cannot be carried out by other experimenters and in other experimental contexts will, in the end be unproductive in science. (Sarkar and Pfeifer 2006 , p.270)

We now turn to an article on “Elevated Plus Maze for Mice” that is available for free on the journal website ( http://www.jove.com/video/1088/elevated-plus-maze-for-mice ). The purpose of this experiment was to investigate anxiety levels in mice through behavioral analysis. The journal article consists of a 9-min video accompanied by text. The video illustrates the handling of the mice in soundproof location with less light, worksheets with characteristics of mice, computer software, apparatus, resources, setting up the computer software, and the video recording of mouse behavior on the computer. The authors describe the apparatus that is used in the experiment and state how procedural differences exist between research groups that lead to difficulties in the interpretation of results:

The apparatus consists of open arms and closed arms, crossed in the middle perpendicularly to each other, and a center area. Mice are given access to all of the arms and are allowed to move freely between them. The number of entries into the open arms and the time spent in the open arms are used as indices of open space-induced anxiety in mice. Unfortunately, the procedural differences that exist between laboratories make it difficult to duplicate and compare results among laboratories.

The authors’ emphasis on the particularity of procedural context echoes in the observations of some philosophers of science:

It is not just the knowledge of experimental objects and phenomena but also their actual existence and occurrence that prove to be dependent on specific, productive interventions by the experimenters” (Sarkar and Pfeifer 2006 , pp. 270-271)

The inclusion of a video of the experimental procedure specifies what the apparatus looks like (Fig.  4 ) and how the behavior of the mice is captured through video recording that feeds into a computer (Fig.  5 ). Subsequently, a computer software which captures different variables such as the distance traveled, the number of entries, and the time spent on each arm of the apparatus. Here, there is visual information at different levels of representation ranging from reconfiguration of raw video data to representations that analyze the data around the variables in question (Fig.  6 ). The practice of levels of visual representations is not particular to the biological sciences. For instance, they are commonplace in nanotechnological practices:

Visual illustration of apparatus

Video processing of experimental set-up

Computer software for video input and variable recording

In the visualization processes, instruments are needed that can register the nanoscale and provide raw data, which needs to be transformed into images. Some Imaging Techniques have software incorporated already where this transformation automatically takes place, providing raw images. Raw data must be translated through the use of Graphic Software and software is also used for the further manipulation of images to highlight what is of interest to capture the (inferred) phenomena -- and to capture the reader. There are two levels of choice: Scientists have to choose which imaging technique and embedded software to use for the job at hand, and they will then have to follow the structure of the software. Within such software, there are explicit choices for the scientists, e.g. about colour coding, and ways of sharpening images. (Ruivenkamp and Rip 2010 , pp.14–15)

On the text that accompanies the video, the authors highlight the role of visualization in their experiment:

Visualization of the protocol will promote better understanding of the details of the entire experimental procedure, allowing for standardization of the protocols used in different laboratories and comparisons of the behavioral phenotypes of various strains of mutant mice assessed using this test.

The software that takes the video data and transforms it into various representations allows the researchers to collect data on mouse behavior more reliably. For instance, the distance traveled across the arms of the apparatus or the time spent on each arm would have been difficult to observe and record precisely. A further aspect to note is how the visualization of the experiment facilitates control of bias. The authors illustrate how the olfactory bias between experimental procedures carried on mice in sequence is avoided by cleaning the equipment.

Our discussion highlights the role of visualization in science, particularly with respect to presenting visualization as part of the scientific practices. We have used case studies from the history of science highlighting a scientist’s account of how visualization played a role in the discovery of DNA and the magnetic field and from a contemporary illustration of a science journal’s practices in incorporating visualization as a way to communicate new findings and methodologies. Our implicit aim in drawing from these case studies was the need to align science education with scientific practices, particularly in terms of how visual representations, stable or dynamic, can engage students in the processes of science and not only to be used as tools for cognitive development in science. Our approach was guided by the notion of “knowledge-as-practice” as advanced by Knorr Cetina ( 1999 ) who studied scientists and characterized their knowledge as practice, a characterization which shifts focus away from ideas inside scientists’ minds to practices that are cultural and deeply contextualized within fields of science. She suggests that people working together can be examined as epistemic cultures whose collective knowledge exists as practice.

It is important to stress, however, that visual representations are not used in isolation, but are supported by other types of evidence as well, or other theories (i.e., in order to understand the helical form of DNA, or the structure, chemistry knowledge was needed). More importantly, this finding can also have implications when teaching science as argument (e.g., Erduran and Jimenez-Aleixandre 2008 ), since the verbal evidence used in the science classroom to maintain an argument could be supported by visual evidence (either a model, representation, image, graph, etc.). For example, in a group of students discussing the outcomes of an introduced species in an ecosystem, pictures of the species and the ecosystem over time, and videos showing the changes in the ecosystem, and the special characteristics of the different species could serve as visual evidence to help the students support their arguments (Evagorou et al. 2012 ). Therefore, an important implication for the teaching of science is the use of visual representations as evidence in the science curriculum as part of knowledge production. Even though studies in the area of science education have focused on the use of models and modeling as a way to support students in the learning of science (Dori et al. 2003 ; Lehrer and Schauble 2012 ; Mendonça and Justi 2013 ; Papaevripidou et al. 2007 ) or on the use of images (i.e., Korfiatis et al. 2003 ), with the term using visuals as evidence, we refer to the collection of all forms of visuals and the processes involved.

Another aspect that was identified through the case studies is that of the visual reasoning (an integral part of Faraday’s investigations). Both the verbalization and visualization were part of the process of generating new knowledge (Gooding 2006 ). Even today, most of the textbooks use the lines of force (or just field lines) as a geometrical representation of field, and the number of field lines is connected to the quantity of flux. Often, the textbooks use the same kind of visual imagery than in what is used by scientists. However, when using images, only certain aspects or features of the phenomena or data are captured or highlighted, and often in tacit ways. Especially in textbooks, the process of producing the image is not presented and instead only the product—image—is left. This could easily lead to an idea of images (i.e., photos, graphs, visual model) being just representations of knowledge and, in the worse case, misinterpreted representations of knowledge as the results of Pocovi and Finlay ( 2002 ) in case of electric field lines show. In order to avoid this, the teachers should be able to explain how the images are produced (what features of phenomena or data the images captures, on what ground the features are chosen to that image, and what features are omitted); in this way, the role of visualization in knowledge production can be made “visible” to students by engaging them in the process of visualization.

The implication of these norms for science teaching and learning is numerous. The classroom contexts can model the generation, sharing and evaluation of evidence, and experimental procedures carried out by students, thereby promoting not only some contemporary cultural norms in scientific practice but also enabling the learning of criteria, standards, and heuristics that scientists use in making decisions on scientific methods. As we have demonstrated with the three case studies, visual representations are part of the process of knowledge growth and communication in science, as demonstrated with two examples from the history of science and an example from current scientific practices. Additionally, visual information, especially with the use of technology is a part of students’ everyday lives. Therefore, we suggest making use of students’ knowledge and technological skills (i.e., how to produce their own videos showing their experimental method or how to identify or provide appropriate visual evidence for a given topic), in order to teach them the aspects of the nature of science that are often neglected both in the history of science and the design of curriculum. Specifically, what we suggest in this paper is that students should actively engage in visualization processes in order to appreciate the diverse nature of doing science and engage in authentic scientific practices.

However, as a word of caution, we need to distinguish the products and processes involved in visualization practices in science:

If one considers scientific representations and the ways in which they can foster or thwart our understanding, it is clear that a mere object approach, which would devote all attention to the representation as a free-standing product of scientific labor, is inadequate. What is needed is a process approach: each visual representation should be linked with its context of production (Pauwels 2006 , p.21).

The aforementioned suggests that the emphasis in visualization should shift from cognitive understanding—using the products of science to understand the content—to engaging in the processes of visualization. Therefore, an implication for the teaching of science includes designing curriculum materials and learning environments that create a social and epistemic context and invite students to engage in the practice of visualization as evidence, reasoning, experimental procedure, or a means of communication (as presented in the three case studies) and reflect on these practices (Ryu et al. 2015 ).

Finally, a question that arises from including visualization in science education, as well as from including scientific practices in science education is whether teachers themselves are prepared to include them as part of their teaching (Bybee 2014 ). Teacher preparation programs and teacher education have been critiqued, studied, and rethought since the time they emerged (Cochran-Smith 2004 ). Despite the years of history in teacher training and teacher education, the debate about initial teacher training and its content still pertains in our community and in policy circles (Cochran-Smith 2004 ; Conway et al. 2009 ). In the last decades, the debate has shifted from a behavioral view of learning and teaching to a learning problem—focusing on that way not only on teachers’ knowledge, skills, and beliefs but also on making the connection of the aforementioned with how and if pupils learn (Cochran-Smith 2004 ). The Science Education in Europe report recommended that “Good quality teachers, with up-to-date knowledge and skills, are the foundation of any system of formal science education” (Osborne and Dillon 2008 , p.9).

However, questions such as what should be the emphasis on pre-service and in-service science teacher training, especially with the new emphasis on scientific practices, still remain unanswered. As Bybee ( 2014 ) argues, starting from the new emphasis on scientific practices in the NGSS, we should consider teacher preparation programs “that would provide undergraduates opportunities to learn the science content and practices in contexts that would be aligned with their future work as teachers” (p.218). Therefore, engaging pre- and in-service teachers in visualization as a scientific practice should be one of the purposes of teacher preparation programs.

Achieve. (2013). The next generation science standards (pp. 1–3). Retrieved from http://www.nextgenscience.org/ .

Google Scholar  

Barber, J, Pearson, D, & Cervetti, G. (2006). Seeds of science/roots of reading . California: The Regents of the University of California.

Bungum, B. (2008). Images of physics: an explorative study of the changing character of visual images in Norwegian physics textbooks. NorDiNa, 4 (2), 132–141.

Bybee, RW. (2014). NGSS and the next generation of science teachers. Journal of Science Teacher Education, 25 (2), 211–221. doi: 10.1007/s10972-014-9381-4 .

Article   Google Scholar  

Chambers, D. (1983). Stereotypic images of the scientist: the draw-a-scientist test. Science Education, 67 (2), 255–265.

Cochran-Smith, M. (2004). The problem of teacher education. Journal of Teacher Education, 55 (4), 295–299. doi: 10.1177/0022487104268057 .

Conway, PF, Murphy, R, & Rath, A. (2009). Learning to teach and its implications for the continuum of teacher education: a nine-country cross-national study .

Crick, F. (1988). What a mad pursuit . USA: Basic Books.

Dimopoulos, K, Koulaidis, V, & Sklaveniti, S. (2003). Towards an analysis of visual images in school science textbooks and press articles about science and technology. Research in Science Education, 33 , 189–216.

Dori, YJ, Tal, RT, & Tsaushu, M. (2003). Teaching biotechnology through case studies—can we improve higher order thinking skills of nonscience majors? Science Education, 87 (6), 767–793. doi: 10.1002/sce.10081 .

Duschl, RA, & Bybee, RW. (2014). Planning and carrying out investigations: an entry to learning and to teacher professional development around NGSS science and engineering practices. International Journal of STEM Education, 1 (1), 12. doi: 10.1186/s40594-014-0012-6 .

Duschl, R., Schweingruber, H. A., & Shouse, A. (2008). Taking science to school . Washington DC: National Academies Press.

Erduran, S, & Jimenez-Aleixandre, MP (Eds.). (2008). Argumentation in science education: perspectives from classroom-based research . Dordrecht: Springer.

Eurydice. (2012). Developing key competencies at school in Europe: challenges and opportunities for policy – 2011/12 (pp. 1–72).

Evagorou, M, Jimenez-Aleixandre, MP, & Osborne, J. (2012). “Should we kill the grey squirrels?” A study exploring students’ justifications and decision-making. International Journal of Science Education, 34 (3), 401–428. doi: 10.1080/09500693.2011.619211 .

Faraday, M. (1852a). Experimental researches in electricity. – Twenty-eighth series. Philosophical Transactions of the Royal Society of London, 142 , 25–56.

Faraday, M. (1852b). Experimental researches in electricity. – Twenty-ninth series. Philosophical Transactions of the Royal Society of London, 142 , 137–159.

Gilbert, JK. (2010). The role of visual representations in the learning and teaching of science: an introduction (pp. 1–19).

Gilbert, J., Reiner, M. & Nakhleh, M. (2008). Visualization: theory and practice in science education . Dordrecht, The Netherlands: Springer.

Gooding, D. (2006). From phenomenology to field theory: Faraday’s visual reasoning. Perspectives on Science, 14 (1), 40–65.

Gooding, D, Pinch, T, & Schaffer, S (Eds.). (1993). The uses of experiment: studies in the natural sciences . Cambridge: Cambridge University Press.

Hogan, K, & Maglienti, M. (2001). Comparing the epistemological underpinnings of students’ and scientists’ reasoning about conclusions. Journal of Research in Science Teaching, 38 (6), 663–687.

Knorr Cetina, K. (1999). Epistemic cultures: how the sciences make knowledge . Cambridge: Harvard University Press.

Korfiatis, KJ, Stamou, AG, & Paraskevopoulos, S. (2003). Images of nature in Greek primary school textbooks. Science Education, 88 (1), 72–89. doi: 10.1002/sce.10133 .

Latour, B. (2011). Visualisation and cognition: drawing things together (pp. 1–32).

Latour, B, & Woolgar, S. (1979). Laboratory life: the construction of scientific facts . Princeton: Princeton University Press.

Lehrer, R, & Schauble, L. (2012). Seeding evolutionary thinking by engaging children in modeling its foundations. Science Education, 96 (4), 701–724. doi: 10.1002/sce.20475 .

Longino, H. E. (2002). The fate of knowledge . Princeton: Princeton University Press.

Lynch, M. (2006). The production of scientific images: vision and re-vision in the history, philosophy, and sociology of science. In L Pauwels (Ed.), Visual cultures of science: rethinking representational practices in knowledge building and science communication (pp. 26–40). Lebanon, NH: Darthmouth College Press.

Lynch, M. & S. Y. Edgerton Jr. (1988). ‘Aesthetic and digital image processing representational craft in contemporary astronomy’, in G. Fyfe & J. Law (eds), Picturing Power; Visual Depictions and Social Relations (London, Routledge): 184 – 220.

Mendonça, PCC, & Justi, R. (2013). An instrument for analyzing arguments produced in modeling-based chemistry lessons. Journal of Research in Science Teaching, 51 (2), 192–218. doi: 10.1002/tea.21133 .

National Research Council (2000). Inquiry and the national science education standards . Washington DC: National Academies Press.

National Research Council (2012). A framework for K-12 science education . Washington DC: National Academies Press.

Nersessian, NJ. (1984). Faraday to Einstein: constructing meaning in scientific theories . Dordrecht: Martinus Nijhoff Publishers.

Book   Google Scholar  

Nersessian, NJ. (1992). How do scientists think? Capturing the dynamics of conceptual change in science. In RN Giere (Ed.), Cognitive Models of Science (pp. 3–45). Minneapolis: University of Minnesota Press.

Nersessian, NJ. (2008). Creating scientific concepts . Cambridge: The MIT Press.

Osborne, J. (2014). Teaching scientific practices: meeting the challenge of change. Journal of Science Teacher Education, 25 (2), 177–196. doi: 10.1007/s10972-014-9384-1 .

Osborne, J. & Dillon, J. (2008). Science education in Europe: critical reflections . London: Nuffield Foundation.

Papaevripidou, M, Constantinou, CP, & Zacharia, ZC. (2007). Modeling complex marine ecosystems: an investigation of two teaching approaches with fifth graders. Journal of Computer Assisted Learning, 23 (2), 145–157. doi: 10.1111/j.1365-2729.2006.00217.x .

Pauwels, L. (2006). A theoretical framework for assessing visual representational practices in knowledge building and science communications. In L Pauwels (Ed.), Visual cultures of science: rethinking representational practices in knowledge building and science communication (pp. 1–25). Lebanon, NH: Darthmouth College Press.

Philips, L., Norris, S. & McNab, J. (2010). Visualization in mathematics, reading and science education . Dordrecht, The Netherlands: Springer.

Pocovi, MC, & Finlay, F. (2002). Lines of force: Faraday’s and students’ views. Science & Education, 11 , 459–474.

Richards, A. (2003). Argument and authority in the visual representations of science. Technical Communication Quarterly, 12 (2), 183–206. doi: 10.1207/s15427625tcq1202_3 .

Rothbart, D. (1997). Explaining the growth of scientific knowledge: metaphors, models and meaning . Lewiston, NY: Mellen Press.

Ruivenkamp, M, & Rip, A. (2010). Visualizing the invisible nanoscale study: visualization practices in nanotechnology community of practice. Science Studies, 23 (1), 3–36.

Ryu, S, Han, Y, & Paik, S-H. (2015). Understanding co-development of conceptual and epistemic understanding through modeling practices with mobile internet. Journal of Science Education and Technology, 24 (2-3), 330–355. doi: 10.1007/s10956-014-9545-1 .

Sarkar, S, & Pfeifer, J. (2006). The philosophy of science, chapter on experimentation (Vol. 1, A-M). New York: Taylor & Francis.

Schwartz, RS, Lederman, NG, & Abd-el-Khalick, F. (2012). A series of misrepresentations: a response to Allchin’s whole approach to assessing nature of science understandings. Science Education, 96 (4), 685–692. doi: 10.1002/sce.21013 .

Schwarz, CV, Reiser, BJ, Davis, EA, Kenyon, L, Achér, A, Fortus, D, et al. (2009). Developing a learning progression for scientific modeling: making scientific modeling accessible and meaningful for learners. Journal of Research in Science Teaching, 46 (6), 632–654. doi: 10.1002/tea.20311 .

Watson, J. (1968). The Double Helix: a personal account of the discovery of the structure of DNA . New York: Scribner.

Watson, J, & Berry, A. (2004). DNA: the secret of life . New York: Alfred A. Knopf.

Wickman, PO. (2004). The practical epistemologies of the classroom: a study of laboratory work. Science Education, 88 , 325–344.

Wu, HK, & Shah, P. (2004). Exploring visuospatial thinking in chemistry learning. Science Education, 88 (3), 465–492. doi: 10.1002/sce.10126 .

Download references

Acknowledgements

The authors would like to acknowledge all reviewers for their valuable comments that have helped us improve the manuscript.

Author information

Authors and affiliations.

University of Nicosia, 46, Makedonitissa Avenue, Egkomi, 1700, Nicosia, Cyprus

Maria Evagorou

University of Limerick, Limerick, Ireland

Sibel Erduran

University of Tampere, Tampere, Finland

Terhi Mäntylä

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Maria Evagorou .

Additional information

Competing interests.

The authors declare that they have no competing interests.

Authors’ contributions

ME carried out the introductory literature review, the analysis of the first case study, and drafted the manuscript. SE carried out the analysis of the third case study and contributed towards the “Conclusions” section of the manuscript. TM carried out the second case study. All authors read and approved the final manuscript.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( https://creativecommons.org/licenses/by/4.0 ), which permits use, duplication, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and permissions

About this article

Cite this article.

Evagorou, M., Erduran, S. & Mäntylä, T. The role of visual representations in scientific practices: from conceptual understanding and knowledge generation to ‘seeing’ how science works. IJ STEM Ed 2 , 11 (2015). https://doi.org/10.1186/s40594-015-0024-x

Download citation

Received : 29 September 2014

Accepted : 16 May 2015

Published : 19 July 2015

DOI : https://doi.org/10.1186/s40594-015-0024-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Visual representations
• Epistemic practices
• Science learning

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Front Psychol

The concepts of representation and information in explanatory theories of human behavior

Focusing in experimental study of human behavior, this article discusses the concepts of information and mental representation aiming the integration of their biological, computational, and semantic aspects. Assuming that the objective of any communication process is ultimately to modify the receiver’s state, the term correlational information is proposed as a measure of how changes occurring in external world correlate with changes occurring inside an individual. Mental representations are conceptualized as a special case of information processing in which correlational information is received, recorded, but also modified by a complex emergent process of associating new elements. In humans, the acquisition of information and creation of mental representations occurs in a two-step process. First, a sufficiently complex brain structure is necessary to establishing internal states capable to co-vary with external events. Second, the validity or meaning of these representations must be gradually achieved by confronting them with the environment. This contextualization can be considered as part of the process of ascribing meaning to information and representations. The hypothesis introduced here is that the sophisticated psychological constructs classically associated with the concept of mental representation are essentially of the same nature of simple interactive behaviors. The capacity of generating elaborated mental phenomena like beliefs and desires emerges gradually during evolution and, in a given individual, by learning and social interaction.

INTRODUCTION

The construction of comprehensive explanatory models of human behavior requires constant review and improvement of concepts in order to integrate different types of structures and levels of implementation. In this sense, this article discuss two concepts frequently used for modeling different aspects of human behavior in biological, psychological, philosophical, physical, and computational explanatory theories. They are the concepts of information and representation. The objective is to discuss the interdependency between both constructs with special attention to their use in experimental investigations of cognitive phenomena.

Briefly, the idea of representation discussed here is related to the brain’s capacity of developing inner states, in the form of relatively stable patterns of neuronal activity, that keep some kind of relationship with events occurring in external world. In many cases, these representations start by simple reactions to external stimuli but, due to brain’s organizational characteristics, evolve by incorporating many other elements than those directly apprehensible from the direct contact with the environment. This capacity of constructing complex mental representations results from a long evolutionary process but its basic constituents can be identified in the neuronal activity of simpler organisms in the form of reactive or conditioned behaviors, for example.

The concept of mental representation in cognitive sciences is frequently associated to complex phenomena such as beliefs and desires. This class of models, also known as representational theories of mind (RTM), consider that these states have “intentionality” in the sense that they are about or refer to things, and may be evaluated with respect to properties like consistency, truth, appropriateness, and accuracy ( Cummins, 1989 ).

This article proposes that the general idea of intentionality or the propriety of mental states of maintaining a correlation with external events can be generalized to describe even the early stages of information processing in the nervous system. This mechanism of co-variation, in association with memory resources and the capacity of generating brain states related to abstract elements of world (more specifically the capacity of deduce the rules governing the behavior of external elements) allow the emergence of the characteristically human cognitive traits.

This broad idea of intentionality is based in a peculiar concept of information as a linking element between brains and world. Information, as used in neurobiological research, can be described as something intrinsically linked to the construction of representations but at the same time as a concept not exclusive of mental instance. Information seems to exist in natural world and human mind has a very special capability of extracting, processing, and using it to increase its capacity of interaction with the environment.

Although frequently studied separately, the concepts of information and representation can be described as having computational and semantic aspects. The term computational refers to the possibility of codification, quantification, manipulation, and physical implementation of information and representations while the term semantic refers to the meaning of both concepts in different contexts.

Information and representation will be discussed here from a neurobiological point of view but with the intention of maintaining coherence with their conceptualization in computational or artificial models of cognition. This coherence requires considering mental representations as biological phenomena, proper but not exclusive of human minds, which construction is achieved by a mechanism of information exchange with the external world. As we shall see below, although representations can be localized in the brain, their meaning does not reside exclusively in the neurobiological instance being a characteristic of the dynamic interaction between brains and environment.

In the following sections, the concepts of mental representation and information will be discussed with a declared bias toward its application in empirical problems of cognitive neurosciences. The interest in these concepts, however, is not restrict to the study of human cognition. Comprehensive discussions about classic information theory can be found in Shannon (1948) , Karnani et al. (2009) , Wang and Shen (2011) , and Adami (2012) . The nature of mental representations in philosophy, psychology, and neurosciences is discussed by Cummins (1989 , 1996 ), Stich (1992) , and Fodor (2000) . Comprehensive discussions about semantic information are found in Floridi (2005) , Karnani et al. (2009) , Jensen et al. (2013) , and Vakarelov (2014) .

THE EMPIRICAL STUDY OF HUMAN BEHAVIOR

The paradigmatic situation faced by neuroscientists during their experimental work can be described as follows: consider an individual observing an object and/or carrying out a mental task while his/her brain activity is recorded by a functional neuroimaging machine. Based in the machine’s outputs, the scientist controlling the experimental setup wants to know how the individual’s cognition works and to what extent the machine output reflects the individual experience of thinking.

Although it is possible to get some kind of information from the machine, the descriptive capacity of this paradigm is limited, especially in relation to the apprehensibility of subjective experiences. This limitation can be expressed by the qualia argument: although the scientist can know something about the individual’s internal state it is impossible for an external observer to have access to the very nature of mental processes because they involve a special quality of conscious experience that cannot be reduced to a linguistically mediated set of descriptive elements ( Kanai and Tsuchiya, 2012 ; Ramos, 2012 ).

This problem can be partially reduced by questioning the individual about her/his subjective experience and checking the accuracy of her/his representations of the external world. This method, however, is limited by the capacity of the individual in accessing their own internal states. The extra-conscious character of many brain activities makes it impossible for someone to be aware and report all elements composing the cognitive experience. Even simple activities are subject to uncontrollable perceptive distortions (optical illusions, for example), spontaneous evocation of memory contents, and subtle affective states that are not consciously perceived.

Although neuroimaging techniques do not account for the qualia question, they are continually improving their capacity of detecting details of brain function in terms of anatomic location and time course of events. The information obtained by functional imaging machines is expressed in terms of electrical signals or measures of metabolic activity which must be articulated with the individual’s linguistic descriptions. Machine recordings allow the spatial localization of structures working at a given moment as well as mapping the time dynamics of their interaction ( Nunez et al., 2014 ). Thus, functional data are collected and analyzed based in a general conceptualization of the brain as an information processing device constituted by specialized and widely interconnected substructures working in constant communication.

INFORMATION BASED ON RECEIVER

Probably, the most influential theory of information is that proposed by Shannon (1948) based in the concept of entropy or the uncertainty associated to the occurrence of a message. The general communication system proposed by Shannon is shown in Figure ​ Figure1 1 .

Information system proposed by Shannon. The red square delimits the receiver component of the system where the information’s semantic value is determined.

In a simplified form, this definition is based on the probability of occurrence of a given message among other possible ones. Although widely explored in computer sciences as well as in the study of interactions between neurons and cortical areas ( Bezzi, 2007 ; Ward and Mazaheri, 2008 ), this approach is not suitable for many other applications in neurosciences. An accurate estimation of the message probability requires previous knowledge of how many other possible messages can possibly occur, which is frequently inaccessible in behavioral studies. In addition, Shannon’s model explicitly does not take in consideration the meaning of the message emitted, transmitted, and received.

The question of meaning of information, centrally important in neurosciences, has been discussed under the general topic of semantic information. Despite the lack of consensus about its definition, semantic information can be described as the data and its meaning, including or not the conditions of truthfulness ( Vakarelov, 2014 ). The study of semantic information has focused in a number of problems, most of them systematized by Floridi (2005) . The main question related to the semantic aspect of information of particular interest for this discussion is “how can data acquire their meaning” ( Floridi, 2005 ; Vakarelov, 2010 ). Vakarelov (2010) , for example, suggests a “pragmatic approach to information where one defines the notion of information system as a special kind of purposeful system emerging within the underlying dynamics of the world, and defines semantic information as the currency of the system. In this way, systems operating with semantic information can be viewed as patterns in organized systems.”

Returning to the general framework of Shannon’s communication system, one can says that the information transmission process is not dependent on the meaning of the message only until reaching the receiver component of the communication system. It occurs because the objective of sending a message is ultimately to provoke changes in receiver’s state. These changes are what determine the existence of the message from the receiver’s point of view. For example, let’s consider an individual in a dark cave populated by bats. In the absence of light and without the capacity of perceiving ultrasounds, the individual can construct only a very partial representation of the cave environment. He/she cannot determine how many bats are inside the cave, what they are doing, and if they are communicating with each other. The observer’s state cannot be modified by the events occurring in the cave due to the absence of adequate sensory mechanisms. For the bats, however, the same environment is full of meaningful information due to their capacity of emitting high frequency sounds and analyzing its echoes. If this individual is a scientist interested in understanding bat behavior, he/she can develop instruments to detect ultrasounds otherwise unperceivable and “extract” more information from the environment. Even with this new instrument, the “meaning” of this new information is not immediately clear. The only way to construct a comprehensible picture of bat activities is by establishing correlations between observable behaviors and the signals obtained by the machine. Although it is impossible for the scientist to get full access to the bat’s mind and to know how is to be like a bat, he/she can map the modifications observed in the environment and compare them with the modifications occurring in the machine states. If the machine is sufficiently precise and the bat’s behavior is sufficiently sophisticated, the scientist can build a limited map of bat’s mind.

This example can be extended to the neuroimaging techniques in general. In brain functional studies the strategy of simply correlating stationary brain states with static external stimuli has been proved meaningless. The simple mapping of all neurons firing at the moment that a specific stimulus is presented does not guarantee that the neural activity observed is related to that act of observation. In order to determine the correlation level between external world and internal brain activity, the strategy is to induce changes in object’s characteristics and observe the resulting changes occurring in brain activity. In functional brain techniques, co-varying patterns of brain activities and object presentation are usually obtained through several repetitions of stereotyped tasks which results are submitted statistical analysis. In fact, the term stimulus used in biological research can be defined as any modification of the environment that interferes with the organism’s state. In this situation, the scientist can check if the observer’s brain is receiving information by identifying changes in neural activity that correlate with changes occurring in the external world.

Therefore, the process that defines the information as something significant occurs in the receiver component of the system (the red box in Figure ​ Figure1 1 ). It does not mean that other components are not relevant but the hypothesis to be explored in the next sections is that the meaning of message emerges in the receiver and any other stimuli running through the information system that is not recognized or that does not induces modifications in the receiver’s state is not information.

The Shannon communication system model has been applied in modeling each step of the nervous system’s functioning. External stimuli work as an information source to sensory cells that generate action potentials and excite the next neuron in the pathway. Cortical areas work as transmitters and receivers in relation to other areas and one person can also be modeled as transmitter, receiver, noise source, or information media according to the interest of the model. Thus, the limits of each component of an information system in an organism are arbitrary and the same formalism used to describe the interaction between two neurons can, in principle, be applied to describe the interactions between neuron nuclei or even between individuals in social interaction.

DEVELOPING REPRESENTATIONS

The co-variation of an observer’s neural/mental states with changes occurring in the external world is the first condition for establishing a representation of objects. Many forms of representation can be generated by this process and several of them may be incomplete or inaccurate. The construction of a set of valid and useful representations requires a complementary mechanism of validation and improvement that, in biological organisms, can be implemented by the process of natural selection.

Tononi (2008) suggests that “through natural selection, epigenesis, and learning, informational relationships in the world mold informational relationships within the main complex that “resonate” best on a commensurate spatial and temporal scale. Moreover, over time these relationships will be shaped by an organism’s values, to refiect relevance for survival. This process can be envisioned as the experiential analog of natural selection. As is well known, selective processes act on organisms through differential survival to modify gene frequencies (genotype), which in turn leads to the evolution of certain body forms and behaviors (extrinsic phenotype).”

Thus, the acquisition of information and creation of mental representations occurs in a two-step process. First, a sufficiently complex brain structure is necessary to establishing internal states capable to co-vary with external events. Second, the validity of these representations must be gradually achieved by confronting them with the environment. The hypothesis discussed here is that the sophisticated psychological constructs classically associated with the concept of mental representation start from simple interactive behaviors. The capacity of using language and interacting in social groups allowed the gradual emergence of more complex human mental phenomena. This development can had occurred even by a relatively disorganized process of creation, modification, and correction of internal states in function of new inputs from external world.

Therefore, it is possible to admit that the mechanisms by which human cognition had developed are present in other classes of organisms. For example, an insect survives in its natural habitat because it can maintain a sufficiently accurate representation of external world. This representation-mediated “world-insect relationship” is limited and it even may not be considered as of cognitive nature. However, the quality and precision of this representation is the optimized result of a compromise between anatomo-physiological constraints and the necessity of providing information processing resources in the context of selective pressure in a specific niche. Partial representations may be suited to improve survival chances because they are easier to be created and corrected and faster to be implemented in natural life situations.

REPRESENTING RULES

Another representational strategy that emerged along the evolution is the representation of the rules or patterns governing what happen in the external world. For example, conditioned behaviors in several animal species can be understood as a representation of external regularities. The increased dog’s salivation after a conditioned stimulus related to food is mediated by a representation, established by learning, of a rule of correlation between two events.

In the human brain, similar mechanisms seem to work even in more complex activities. Noelle (2012) reviewed evidences that rule-guided behaviors in humans are associated with the functioning of the prefrontal cortex, the basal ganglia, and related brain structures. The author suggests that a “dopamine-based gating mechanism interacts with standard models of synaptic plasticity to support the development of appropriately isolated and dimensional prefrontal representations, giving rise to improved generalization to novel situations when adequately diverse training experiences are provided.” According to this proposal, some regions of the prefrontal cortex may encode references or “pointers” to other prefrontal areas in a representational scheme that would allow for essentially combinatoric generalization to novel rules. This capacity of combinatoric generalization does not imply a “mere implementation” of symbolic rule-interpretation mechanisms. For Noelle, “complex interactions between the rule representations actively maintained in prefrontal cortex and the dynamic processes of more posterior neural circuits give rise to graded and context-sensitive patterns of performance that escape description by a purely symbolic rule account. Also, statistical regularities in the experiences present during the development of prefrontal cortex can profoundly shape the kinds the explicit rules that can robustly be represented and applied.”

The process of information processing based on representation of rules can be further enhanced by the creation of subsets of a priori representations available for use in natural situations. Innate behaviors, related to threat detection for example, require the pre-existence of relatively complex representations capable of enhancing fast protective actions. This characteristic is called preparedness of fear and phobias and it has been identified also in human behavior. Mineka and Ohman (2002) present evidences for the existence of an evolved module for fear elicitation and fear learning with four primary characteristics: “First, it is preferentially activated by stimuli related to survival threats in evolutionary history. Thus, fear-relevant stimuli lead to superior conditioning of aversive associations compared with fear-irrelevant stimuli. Second, the module is automatically activated by fear-relevant stimuli, meaning that fear activation occurs before conscious cognitive analysis of the stimulus can occur. Third, the fear module is relatively impenetrable to conscious cognitive control, and fear conditioning with fear-relevant stimuli can occur even with subliminal conditioned stimuli. Fourth, the amygdala seems to be the central brain area dedicated to the fear module.”

The high velocity required by the process of identifying threats and implementing adequate responses imply in an increased probability of errors related to the simplification of external situations, misinterpretation of new events, and ultimately the creation of distorted representations. This style of cognitive functioning can be understood under a biological perspective where, in natural situations, errors of commission (wrongly reacting to a non-threat) are more acceptable than errors of omission (not reacting to a real threat).

Other cognitive capacities like empathy and face recognizing also seem to be implemented by similar mechanisms of working with pre-prepared representations ( Regenbogen et al., 2012 ; Kryklywy et al., 2013 ; Prochnow et al., 2013 ). Admitting that the same design strategy is used in the implementation of other cognitive functions, this mechanism of simplifying representations in order to facilitate stimuli responses may be hypothesized as playing a role in complex phenomena associated to partial or biased evaluations of external situations like folk psychological explanations and the occurrence of preconceptions in social contexts.

CORRELATION AND INFORMATION

In order to differentiate from Shannon’s informational entropy, the term correlational information is proposed here, not as a measure of probability but as a measure of how changes occurring in external world correlate with changes occurring inside an agent. This concept does not depend either on the physical, biological, or linguistic nature of external object nor on the cognitive capacity of the receiver. Correlational information depends on the receiver capacity of modifying aspects of its internal states in function of changes occurring in the external environment. This receiver’s plasticity needs not to reflect every characteristic of external objects because even partial representations can be sufficient for adequate interactions with the environment. This strategy of adopting an information model based in correlations aims to emphasize the importance of the receiver component of in the general model of information system.

This approach can be illustrated as follow: let’s consider an animal visually perceiving a light source emitting different colors ( Figure ​ Figure2A 2A ). If its sensory organ has the capacity of having its state modified in a given way by each color which induces one corresponding change of state ( Figure ​ Figure2B 2B ), one can says that this animal is capable of having accurate color perception. Note that, in this model, how exactly this correspondence is physically implemented is not important. The central point is that the path blue, red, green, yellow in the external world correspond to the path a,b,c,d inside the organism. In contrast, if the sensory organ is not capable of distinguishing blue from green and red from yellow, for example, its internal representation is given by a simpler path (fig) in Figure ​ Figure2C 2C .

This figure illustrates different representational capacities of sensory organs. (A) represents a light source emitting different colors. The sensory organ illustrated in (B) is capable of associating the series of different states a,b,c,d where each state is related to a different color blue, red, green, and yellow. (C) illustrate another kind of sensory organ not capable of distinguishing blue from green and red from yellow and, therefore, representing the changes occurring in the exterior world by simplified set of states (f,g).

The representation expressed in Figure ​ Figure2C 2C is partial in comparison to that expressed by Figure ​ Figure2B 2B but its physical implementation by a simpler sensory organ demands less resources. If both representations have the same efficiency in preserving the animal’s life (detecting food or predators, for example), the simplest alternative may be the most advantageous unless new changes occur in the environment making the exact color perception an essential trait for survival.

According this model, the flux of correlational information along nervous system is the set of modifications gradually established along sensory cells, nerves, interneurons, and brain structures involved in behavior expression. An advantage of this concept is that these modifications are potentially detectable by functional techniques although not always accessible to an individual’s consciousness. In experimental context, even physiological manifestations like, for example changes in autonomic functioning or postural control can be considered as part of the set of information that composes mental representations. The inclusion of these not purely cognitive elements is essential, for example, in the study of emotions where several experiential elements cannot be adequately described by language.

This proposal does not imply in denying the existence of internally generated states. Although mental events can occur with a degree of independence from external influences (for example, reflections, interpretations, and mathematical thinking) the basic neural components that allowed the development of these sophisticated capacities are closely related to those working in other relatively more simple brain activities.

The human thinking process can run with a relative independence from external inputs like in mental fantasies. The correlational model proposes is that the ability of working at this level of abstraction was acquired by the gradual improvement of the capacity of using correlational information. Once acquired, this ability allows to the individual to work with independence from direct sensorial inputs and add new elements to mental contents. Although fantasies can be generated with large degree of freedom, the awareness that these contents are internally created is given by the capacity of confronting them with external inputs.

One example of internally generated state involving pre-prepared structures closely related with external events is the mirror neurons system. Originally found in macaque monkeys, in the ventral premotor cortex, area F5 and inferior parietal lobule, this group of neurons fire when the animal sees another animal (or the experimenter) performing actions similar to those pertaining to its natural repertoire of actions. Neuroimaging and electrophysiological studies indicate that mirror neurons may serve for action recognition in monkeys as well as humans, whereas their putative role in imitation and language may be realized in human but not in monkey ( Oztop et al., 2013 ). Although primarily of motor nature, mirror neurons have been associated with mental activities like intention understanding, emotions, empathy, and speech ( Acharya and Shukla, 2012 ).

Another examples of mental representations based in brain-environment co-variant proprieties are those involved in the orientation and movement in the space. Land (2014) points out “that the motor system requires a representation of space that maintains a consistent relationship with objects in the outside world as the body moves within it, then this could also serve as a model of a stable outside world of which we can be conscious. A high-definition representation is not necessary, all that is required is that it provides a stable framework to which detailed information, provided by the visual pathways through the occipital and temporal lobes, can be temporarily attached.”

The creation and recording of mental representations involves the gradual recruiting of relatively distant but highly connected brain components with different time dynamics. Consequently, mental representations are not localized in specific brain regions but they gradually emerge along the entire neuronal processing. This idea is compatible with several neurobiological phenomena associated with conscious experience. Shen et al. (2013) proposed that the experience of “insight,” described as an experience related to a state of understanding, which emerges into one’s conscious awareness with sudden abruptness, involves many distributed brain regions, including the lateral prefrontal cortex, cingulate cortex, hippocampus, superior temporal gyrus, fusiform gyrus, precuneus, cuneus, insula, cerebellum, and some areas of the parietal cortex.

The ability of processing complex concepts and rules governing external events is essential to the emergence of another property of human cognitive systems that is the possibility of anticipating future events. The capacity of preview the occurrence of a given stimulus can be identified even in simple organisms exhibiting conditioned behaviors. For example, the technique of olfactory conditioning of the sting extension response has been extensively used to yield new insights into the rules and mechanisms of aversive learning in insects ( Tedjakumala and Giurfa, 2013 ).

This simple capacity of representing rules can be improved by the development of more complex neural resources. In fact, this capacity vary from one species to other ( Seed et al., 2012 ) and along the cognitive development of each individual ( Wellman et al., 2001 ). Moreover, there are also evidences that this representational capacity do not depend of neuronal mechanisms but also of adequate social and cultural influences ( Moriguchi, 2014 ).

EMERGENCE AND COMPLEXITY

The next question, central for this discussion, is how simple mechanisms of correlation allow the emergence of complex abstractions in the human mind. A possible strategy for clarifying this point is to explore complex systems theories and its applicability at the several structural and organizational levels evolved in the genesis of human behavior.

The idea that complex patterns can spontaneously emerge from simpler components is largely discussed in natural sciences and a number of theoretical ideas have been proposed to explain their occurrence like, for example agent-based models and genetic algorithms ( Caticha and Vicente, 2011 ; Gros, 2013 ).

One of these theoretical models in particular, known as self-organized criticality (SOC), has received great attention as a possible explanation for the spontaneous emergence of complex patterns both at neural and behavioral levels. The concept of SOC was proposed by Bak et al. (1987) as one of the mechanisms by which complexity arises in nature. They suggested that “dissipative dynamical systems with extended degrees of freedom can evolve toward a self-organized critical state, with spatial and temporal power-law scaling behavior.” This spatial scaling leads to self-similar “fractal” structure identifiable in many conditions.

Beggs and Plenz (2003) reported evidences of this phenomenon studying organotypic cultures from coronal slices of rat somatosensory cortex. They continuously recorded spontaneous local field potentials (LFPs) using a 60 channel multielectrode array and found that the propagation of synchronized LFPs activity was described by a power-law. The authors suggested the slope of this power-law, as well as its branching parameter, indicate the presence of SOC in these preparations. ( Beggs and Plenz, 2003 ) found evidence that the critical branching process optimizes information transmission while preserving stability in cortical networks. Simulations showed that a branching parameter at value found in the experimental preparation optimizes information transmission in feed forward networks, while preventing runaway network excitation. The authors called this pattern “neuronal avalanches” and hypothesized that it could be a generic property of cortical networks and represent a mode of activity differing from oscillatory, synchronized, or wave-like network states.

Compatible with the ideas discussed here, the identification of such patterns of functioning seems to depend on the brain functioning in context. El Boustani et al. (2009) studied intracellular activity of 15 neurons in the primary visual cortex of the anesthetized and paralyzed cat. Each neuron was recorded while presenting four full field stimuli through the dominant eye: a drifting grating at the cell’s optimal orientation and spatial frequency, a high spatial definition dense noise, a natural image animated with a simulated eye movement sequence, and a grating animated with the same eye movement sequence. The authors found the recordings displayed power-law frequency scaling at high frequencies, with a fractional exponent dependent on the spatio-temporal statistics of the visual stimuli. They also reported that this effect was reproduced in computational models of a recurrent network. They noted “that the power-law relations found here depend on the stimulus, which means that the frequency scaling exponent does not represent a unique signature of cortical network activity, but rather reflects a measure of the dynamic interplay between the sensory evoked activity and the ongoing recurrent network activity.”

The possibility of SOC being relevant for explaining complex human behavior was explored by Ramos et al. (2011) who evaluated groups of individuals with and without mental disorders in social interaction during several weeks. Although the behavior of each individual had been very different from other participants in absolute terms, the statistical description of the different groups (individuals with depression, psychosis, mania, and normal controls) showed identical patterns of behavioral variation. In all groups, comparing the behavior of individuals with themselves, small changes of behavior were very frequent while large variations were rare. The characteristic of having the same variation pattern reproduced at different levels of human activity, suggests the presence of self-similarity ( Serrano et al., 2008 ). The curves describing the behavior of all clinical groups and controls showed the same aspect and fitted a power-law. The authors suggested that the presence of self-similarity and power-laws is compatible with the hypothesis that humans in social interaction constitute a system exhibiting SOC.

Self-organized criticality is certainly a promising concept for integrating biological and behavioral aspects of human behavior under the same causal mechanisms but it doubtless requires more empirical investigations ( Hidalgo et al., 2014 ).

A BRIEF COMMENT ON THE SEMANTIC QUESTION

The last important point to be discussed here is the question of the ascription of meaning in informational models of cognition. This is a very problematic discussion in the literature that cannot be adequately addressed in the limited scope of this article. However, the empirical research in neurosciences demands some strategy for dealing with this problem due to the impossibility of understanding many aspects of human behavior without considering some form of justification.

A possible provisional strategy is to leave the concept of meaning momentarily aside and explore a utilitarian approach of the mental representations. In a biological perspective, the immediate utility of behaviors and mental representations is increasing individual’s survival chances in different contexts. So, although informations and representations have been defined, in this correlational approach, in function of effects observed in the receiver component, their utilitarian character must be apprehended only in the context of the entire communication system.

Naturally, the idea that human cognition was molded by evolutionary mechanisms is not new. Tononi (2008) explain this hypothesis: “Brain mechanisms, including those inside the main complex, are what they are by virtue of along evolutionary history, individual development, and learning. Evolutionary history leads to the establishment of certain species-specific traits encoded in the genome, including brains and means to interact with the environment. Development and epigenetic processes lead to an appropriate scaffold of anatomical connections. Experience then refines neural connectivity in an ongoing manner though plastic processes, leading to the idiosyncrasies of the individual “connectome” and the memories it embeds.”

The general concepts of evolution theory have been used for the explanation of several kinds of behaviors and cognitive phenomena. However, this explanatory strategy still needs to be better incorporated by empirical studies. The same attention dedicated to developing neurofunctional techniques must also be dedicated to the identification and analysis to the characteristics of the environment where the behaviors are manifested. For example, this utilitarian characteristic of informational models suggests that future developments in functional brain studies must consider the use of immersive virtual reality setups as a way of controlling the behavioral context.

CONCLUDING REMARKS

This article aimed to address some questions about the use of the representation and information concepts in the context of experimental research in cognitive sciences. The focus in the “information based on the receiver” proposed here is justified by the interest of developing objective approaches to the study of human behavior in biological and semantic terms. This search for new conceptual approaches took the risk of being superficial in its formalism but it was proposed as a first step for the description of the different elements that contribute to the construction of mental representations.

The correlational information concept discussed here aimed to be sufficiently simple to allow a naturalization of the information concept in the sense that all interaction between physical entities can be seen as an informational phenomenon. In this model, the construction of mental representations can be seen as a special case of information processing in which correlational information is received, recorded, but also modified by a complex, emergent, and possibly stochastic process of associating new elements. The validity of these new internally generated constituent elements is granted by its continuous confrontation with new external inputs and by the selection of the most adequate representations in relation to its capacity of improving survival chances.

The hypothesis is that this basic mechanism works in all animal species but, with the improved human brain capacity, it leads to the emergence of higher order or abstract descriptive elements of external objects that allow the prediction of future events. This process is possible by the manipulation of internal states representing not only objects but also the rules governing their behavior. In this model, although the content of correlational information depend on the receiver capacity of creating internal states capable to co-vary with external events, the utility of a given information can be apprehended only by the observation of the entire communication system.

The continuous process of collecting information, creating representations, generating predictions, comparing with outcomes, and adjusting them in order to optimize their accuracy is compatible with several psychological models of learning and cognitive development. This mechanism of correlational representations is also compatible with a Bayesian conception of cognitive functioning where partial or provisional representations work as estimators of a priori probabilities in dealing with future events ( Tenenbaum et al., 2011 ).

The ideas discussed here represent a first approach and naturally demand deeper investigations in relation to its theoretical and empirical implications. In theoretical terms, although theories like SOC are promising in explaining human behavior, other mathematical models also deserve attention. Caticha and Vicente (2011) , for example, argue that statistical mechanics can leads to aggregated predictions which can be tested against extensive data sets with partial information about populations. The process of exchanging information and learning patterns involved in these models can elicits collective emergent properties not found in individual behaviors.

In relation to the empirical research, this discussion suggests that the integrative study of the computational and semantic elements that compose human experiences will demand significant technical and theoretical improvements. Technically, the combined register of different variables like cortical electric activity, mapping of eye movements, measures of skin galvanic conductance, and postural control obtained during carefully planned cognitive activities emulated in virtual reality environments, for example, can potentially give a deeper comprehension of the mental, affective, and motor events occurring in realistic contexts.

Conflict of Interest Statement

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

The author is grateful to J. T. S. Rebouças and Ronald C. Ranvaud for insightful discussions and review of early paper’s versions.

• Acharya S., Shukla S. (2012). Mirror neurons: enigma of the metaphysical modular brain. J. Nat. Sci. Biol. Med. 3 118–124 10.4103/0976-9668.101878 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Adami C. (2012). The use of information theory in evolutionary biology. Ann. N. Y. Acad. Sci. 1256 49–56 10.1111/j.1749-6632.2011.06422.x [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Bak P., Tang C., Wiesenfeld K. (1987). Self-organized criticality: an explanation of the 1/f noise. Phys. Rev. Lett. 59 381–384 10.1103/PhysRevLett.59.381 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Beggs J. M., Plenz D. (2003). Neuronal avalanches in neocortical circuits. J. Neurosci. 23 11167–11177 [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Bezzi M. (2007). Quantifying the information transmitted in a single stimulus. Biosystems 89 4–9 10.1016/j.biosystems.2006.04.009 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Caticha N., Vicente R. (2011). Agent-based social psychology: from neurocognitive processes to social data F. Adv. Complex Syst. 14 711–731 10.1142/S0219525911003190 [ CrossRef ] [ Google Scholar ]
• Cummins R. (1989). Meaning and Mental Representation Cambridge, MA: MIT Press [ Google Scholar ]
• Cummins R. (1996). Representations, Targets, and Attitudes. Cambridge, MA: MIT Press [ Google Scholar ]
• El Boustani S., Marre O., Béhuret S., Baudot P., Yger P., Bal T., et al. (2009). Network-state modulation of power-law frequency-scaling in visual cortical neurons. PLoS Comput. Biol. 5 : e1000519 10.1371/journal.pcbi.1000519 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Floridi L. (2005). Is semantic information meaningful data? Philos. Phenomenol. Res. 70 351–370 10.1111/j.1933-1592.2005.tb00531.x [ CrossRef ] [ Google Scholar ]
• Fodor J. A. (2000). The Mind Doesn’t Work That Way : The Scope and Limits of Computational Psychology Cambridge, MA: MIT Press [ Google Scholar ]
• Gros C. (2013). Complex and Adaptive Dynamical Systems : A Primer New York: Springer; 10.1007/978-3-642-36586-7 [ CrossRef ] [ Google Scholar ]
• Hidalgo J., Grilli J., Suweis S., Mu noz M. A., Banavar J. R., Maritan A. (2014). Information-based fitness and the emergence of criticality in living systems. Proc. Natl. Acad. Sci. U.S.A. 111 10095–10100 10.1073/pnas.1319166111 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Jensen G., Ward R., Balsam P. (2013). Information: theory, brain, and behavior. J. Exp. Anal. Behav. 100 408–431 10.1002/jeab.49 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Kanai R., Tsuchiya N. (2012). Qualia. Curr. Biol. 22 R392–R396 10.1016/j.cub.2012.03.033 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Karnani M., Paakkonen K., Annila A. (2009). The physical character of information. Proc. Roy. Soc. Math. Phys. Eng. Sci. 465 2155–2175 10.1098/rspa.2009.0063 [ CrossRef ] [ Google Scholar ]
• Kryklywy J. H., Nantes S. G., Mitchell D. G. (2013). The amygdala encodes level of perceived fear but not emotional ambiguity in visual scenes. Behav. Brain Res. 252 396–404 10.1016/j.bbr.2013.06.010 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Land M. F. (2014). Do we have an internal model of the outside world? Philos. Trans. R. Soc. Lond. B Biol. Sci. 369 20130045 10.1098/rstb.2013.0045 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Mineka S., Ohman A. (2002). Phobias and preparedness: the selective, automatic, and encapsulated nature of fear. Biol. Psychiatry 52 927–937 10.1016/S0006-3223(02)01669-4 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Moriguchi Y. (2014). The early development of executive function and its relation to social interaction: a brief review. Front. Psychol. 5 : 388 10.3389/fpsyg.2014.00388 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Noelle D. C. (2012). On the neural basis of rule-guided behavior. J. Integr. Neurosci. 11 453–475 10.1142/S021963521250029X [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Nunez P. L., Srinivasan R., Fields R. D. (2014). EEG functional connectivity, axon delays and white matter disease. Clin. Neurophysiol. 10.1016/j.clinph.2014.04.003 [Epub ahead of print] [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Oztop E., Kawato M., Arbib M. A. (2013). Mirror neurons: functions, mechanisms and models. Neurosci. Lett. 540 43–55 10.1016/j.neulet.2012.10.005 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Prochnow D., Kossack H., Brunheim S., Müller K., Wittsack H. J., Markowitsch H. J., et al. (2013). Processing of subliminal facial expressions of emotion: a behavioral and fMRI study. Soc. Neurosci. 8 448–461 10.1080/17470919.2013.812536 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ramos R. T. (2012). The conceptual limits of neuroimaging in psychiatric diagnosis. AJOB Neurosci. 3 52–53 10.1080/21507740.2012.721856 [ CrossRef ] [ Google Scholar ]
• Ramos R. T., Sassi R. B., Piqueira J. R. C. (2011). Self-organized criticality and the predictability of human behavior. New Ideas Psychol. 29 38–48 10.1016/j.newideapsych.2009.12.001 [ CrossRef ] [ Google Scholar ]
• Regenbogen C., Schneider D. A., Finkelmeyer A., Kohn N., Derntl B., Kellermann T., et al. (2012). The differential contribution of facial expressions, prosody, and speech content to empathy. Cogn. Emot. 26 995–1014 10.1080/02699931.2011.631296 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Seed A., Seddon E., Greene B., Call J. (2012). Chimpanzee ‘folk physics’: bringing failures into focus. Philos. Trans. R. Soc. Lond. B Biol. Sci. 367 2743–2752 10.1098/rstb.2012.0222 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Serrano M. A., Krioukov D., Bogu ná M. (2008). Self-similarity of complex networks and hidden metric spaces. Phys. Rev. Lett. 100 078701 10.1103/PhysRevLett.100.078701 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Shannon C. E. (1948). A mathematical theory of communication. Bell Syst. Tech. J. 27 379–423 10.1002/j.1538-7305.1948.tb01338.x [ CrossRef ] [ Google Scholar ]
• Shen W. B., Luo J. C. L., Yuan Y. (2013). New advances in the neural correlates of insight: a decade in review of the insightful brain. Chin. Sci. Bull. 58 1497–1511 10.1007/s11434-012-5565-5 [ CrossRef ] [ Google Scholar ]
• Stich S. (1992). What is a theory of mental representation? Mind 101 243–261 10.1093/mind/101.402.243 [ CrossRef ] [ Google Scholar ]
• Tedjakumala S. R., Giurfa M. (2013). Rules and mechanisms of punishment learning in honey bees: the aversive conditioning of the sting extension response. J. Exp. Biol. 216 2985–2997 10.1242/jeb.086629 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Tenenbaum J. B., Kemp C., Griffiths T. L., Goodman N. D. (2011). How to grow a mind: statistics, structure, and abstraction. Science 331 1279–1285 10.1126/science.1192788 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Tononi G. (2008). Consciousness as integrated information: a provisional manifesto. Biol. Bull. 215 216–242 [ PubMed ] [ Google Scholar ]
• Vakarelov O. (2010). Pre-cognitive semantic information. Know. Techn. Pol. 23 193–226 10.1007/s12130-010-9109-5 [ CrossRef ] [ Google Scholar ]
• Vakarelov O. (2014). From interface to correspondence: recovering classical representations in a pragmatic theory of semantic information. Minds Mach. 24 327–351 10.1007/s11023-013-9318-2 [ CrossRef ] [ Google Scholar ]
• Wang C., Shen H. (2011). Information theory in scientific visualization. Entropy 13 254–273 10.3390/e13010254 [ CrossRef ] [ Google Scholar ]
• Ward B., Mazaheri Y. (2008). Information transfer rate in fMRI experiments measured using mutual information theory. J. Neurosci. Methods 167 22–30 10.1016/j.jneumeth.2007.06.027 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Wellman H. M., Cross D., Watson J. (2001). Meta-analysis of theory-of-mind development: the truth about false belief. Child Dev. 72 655–684 10.1111/1467-8624.00304 [ PubMed ] [ CrossRef ] [ Google Scholar ]

Mathematics in Physics Education pp 75–102 Cite as

Mathematical Representations in Physics Lessons

• Marie-Annette Geyer 4 &
• Wiebke Kuske-Janßen 4  
• First Online: 03 July 2019

1070 Accesses

9 Citations

Physics is characterized by the use of specific types of representations. These representations play an important role in the teaching and learning of physics. This chapter/article starts with a general description of representations from a cognitive sciences' and semiotics' view and presents the state of theory about representations in physics education with focus on mathematical ones. Based on this, an adapted classification of representations in physics lessons is presented. This is followed by theoretical considerations and empirical findings about the relevance of (different) representations for learning and understanding physics and students' difficulties in this context. Furthermore, two models are introduced that are part of current research and offer an approach to analyse different changes of representations in physics classes. Ultimately, several implications for teaching will be derived.

The authors contributed equally to this work.

This is a preview of subscription content, log in via an institution .

Buying options

• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Durable hardcover edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Airey and Linder ( 2009 ) use the term semiotic resources and distinguish between tools, activities, and representations. In our argumentation this distinction is not important; tools and activities can be seen predominantly as handling with objective representations.

Fredlund et al. ( 2012 ) use the terminology of semiotic resources as well; therefore they call it the disciplinary affordance of a semiotic resource .

Adu-Gyamfi, K., Stiff, L. V., & Bossé, M. J. (2012). Lost in translation: Examining translation errors associated with mathematical representations. School Science and Mathematics, 112 (3), 159–170.

Article   Google Scholar  

Ainsworth, S. (2006). DeFT: A conceptual framework for considering learning with multiple representations. Learning and Instruction, 16 , 183–198.

Ainsworth, S. (2008). The educational value of multiple-representations when learning complex scientific concepts. In J. K. Gilbert et al. (Eds.), Visualization: theory and practice in science education (pp. 191–208). Dordrecht: Springer.

Chapter   Google Scholar  

Ainsworth, S., Bibby, P. A., & Wood, D. J. (1998). Analysing the costs and benefits of multi-representational learning environments. In M. Someren (Ed.), Learning with multiple representations (pp. 120–134). Pergamon, as cited in Bauer (2015).

Google Scholar  

Airey, J. (2009). Science, language and literacy. Case studies of learning in Swedish University physics . Uppsala: Uppsala University.

Airey, J., & Linder, C. (2009). A Disciplinary Discourse Perspective on University Science Learning: Achieving Fluency in a Critical Constellation of Modes. Journal of Research in Science Teaching, 46 (1), 27–49.

Bagno, E., Eylon, B.-S. & Berger, H. (2011). How to promote the learning of physics from formulae? In: GIREP-EPEC & PHEC 2009 International Conference (=Physics Community and Cooperation Vol. 2), pp. 77–83.

Bauer, A. (2015). Argumentieren mit multiplen und dynamischen Repräsentationen . Würzburg: Würzburg University Press.

Bruner, J. S. (1974). Entwurf einer Unterrichtstheorie . Berlin: Berlin Verlag.

Bossé, M. J., Adu-Gyamfi, K., & Cheetham, M. R. (2011). Assessing the difficulty of mathematical translations: Synthesizing the literature and novel findings. International Electronic Journal of Mathematics Education, 6 (3), 113–133.

Chandler, P. & Sweller, J. (1991). Cognitive load theory and the format of instruction. Cognition and Instruction , 8 (4), 293–332, as cited in Bauer (2015).

diSessa, A. A. (2004). Metarepresentation: Native competence and targets for instruction. Cognition and Instruction, 22 (3), 293–331.

Dolin, J. (2016). Representations in science. In R. Gunstone (Ed.), Encyclopedia of science education (pp. 836–838). Springer Science+Business Media Dordrecht 2015, Corrected Printing 2016.

Duval, R. (2006). A cognitive analysis of problems of comprehension in a learning of mathematics. Educational Studies in Mathematics, 61 (1). Springer, 103–131.

Eriksson, U., Linder, C., Airey, J., & Redfors, A. (2014). Introducing the anatomy of disciplinary discernment: an example from astronomy. European Journal of Science and Mathematics Education, 2 (3), 167–182.

Fredlund, T., Airey, J., & Linder, C. (2012). Exploring the role of physics representations: An illustrative example from students sharing knowledge about refraction. European Journal of Physics, 33 , 657–666.

Geyer, M. –A., & Pospiech, G. (2015). Darstellungen funktionaler Zusammenhänge im Physikunterricht. Darstellungswechsel in der Sekundarstufe 1. In PhyDidB –Didaktik der Physik –Beiträge zur DPG-Frühjahrstagung Wuppertal 2015 .

Gilbert, J. K. (2016). Chapter 7. The contribution of visualisation to modelling-based teaching. In J. K. Gilbert & R. Justi (Eds.), Modelling-based teaching in science education (Models and modeling in science education) (Vol. 9, pp. 121–148). Cham: Springer.

Hattikudur, S., Prather, R. W., Asquith, P., Alibali, M. W., Knuth, E. J., & Nathan, M. (2012). Constructing graphical representations: Middle schoolers’ intuitions and developing knowledge about slope and Y-intercept. School Science and Mathematics, 112 (4), 230–240.

Hoffmann, L. (1987). Kommunikationsmittel Fachsprache. Eine Einführung . 3., durchgesehene Auflage. Berlin: Akademie-Verlag (=Sammlung Akademie-Verlag Bd.44. Sprache).

Janßen, W., & Pospiech, G. (2015). Versprachlichung von Formeln und physikalisches Formelverständnis. In S. Bernholt (Hrsg.), Heterogenität und Diversität – Vielfalt der Voraussetzungen im naturwissenschaftlichen Unterricht. Gesellschaft für Didaktik der Chemie und Physik, Jahrestagung in Bremen 2014 (S. 636–638). Kiel: IPN.

Janvier, C. (1987). Translation process in mathematics education. In Problems of representation in the teaching and learning of mathematics (pp. 27–32). Hillsdale: Lawrence Erlbaum.

Karam, R., & Pietrocola, M. (2010). Recognizing the structural role of mathematics in physical thought. In M. F. Tasar & G. Çakmakci (Eds.), Contemporary science education research: International perspectives (pp. 65–76). Ankara: Pegem Akademi.

Kim, E., & Pak, S.-J. (2002). Students do not overcome conceptual difficulties after solving 1000 traditional problems. American Journal of Physics, 70 (7), 759–765.

Kozma, R. (2003). The material features of multiple representations and their cognitive and social affordances for science understanding. Learning and Instruction, 13 , 205–226.

Kozma, R. B., Russell, J., Jones, T., Marx, N., & Davis, J. (1996). The use of multiple, linked representations to facilitate science understanding. In S. Vosniadou (Ed.),. (Hrsg.) International perspectives on the design of technology-supported learning environments (pp. 41–60). Mahwah: L. Erlbaum Associates, as cited in Bauer (2015).

Leinhardt, G., Zaslavsky, O., & Stein, M. K. (1990). Functions, graphs, and graphing: Tasks, learning, and teaching. Review of Educational Research, 60 (1), 1–64.

Leisen, J. (2004). Der Wechsel der Darstellungsformen als wichtige Strategie beim Lehren und Lernen im DFU. Fremdsprache Deutsch, 30 , 15–21.

Leisen, J. (2005). Wechsel der Darstellungsformen. Ein Unterrichtsprinzip für alle Fächer. Der Fremdsprachliche Unterricht Englisch, 78 , 9–11.

Lemke, J. L. (1998). Multiplying meaning: Visual and verbal semiotics in scientific text. In J. R. Martin & R. Veel (Eds.), Reading science: Critical and functional perspectives on discourse of science (pp. 87–113). London: Routledge.

Lemke, J. L. (2002). Mathematics in the middle: Measure, picture, gesture, sign, and word. In M. Anderson, A. Saenz-Ludlow, S. Zellweger, & V. Cifarelli (Eds.), Educational perspectives on mathematics as semiosis: From thinking to interpreting to knowing (pp. 215–234). Ottawa: Legas Publishing.

Lemke, J. L. (2004). The literacies of science. In E. Wendy Saul (Ed.), Crossing borders in literacy and science instruction (pp. 33–47). Newark, DE/Arlington, VA: International Reading Association/NSTA Press.

Linder, C. (2013). Disciplinary discourse, representation, and appresentation in the teaching and learning of science. European Journal of Science and Mathematics Education, 1 (2), 43–49.

Malle, G., Wittmann, E. C., & Bürger, H. (1993). Didaktische Probleme der elementaren Algebra . Wiesbaden: Springer.

Book   Google Scholar  

Mayer, R. E. & Sims, V. (1994). For whom is a picture worth 1000 words - Extensions of a dual-coding theory of multimedia learning. Journal of Educational Psychology , 86 (3), 389–401, as cited in Bauer (2015 ).

McDermott, L. C., Rosenquist, M. L., & van Zee, E. H. (1987). Student difficulties in connecting graphs and physics: Examples from kinematics. American Journal of Physics, 55 (6), 503–513.

Nitsch, R. (2015). Diagnose von Lernschwierigkeiten im Bereich funktionaler Zusammenhänge. Eine Studie zu typischen Fehlermustern bei Darstellungswechseln . Wiesbaden: Springer Spektrum.

Peirce, C. S. (1960). Collected papers of Charles Sanders Peirce (2. Aufl.). Cambridge, MA: Belknap Press of Harvard University Press.

Pietrocola, M. (2008). Mathematics as structural language of physical thought. In M. Vicentini & E. Sassi (Eds.), Connecting research in physics education with teacher edu-cation , Bd. 2. ICPE.

Planinic, M., Ivanjek, L., Susac, A., & Milin-Sipus, Z. (2013). Comparison of university students’ understanding of graphs in different contexts. Physical Review Special Topics – Physics Education Research, 9 (2), 020103, 1–9.

Pospiech, G., & Oese, E. (2014). The use of mathematical elements in physics – View of grade 8 pupils. In L. Dvořák & V. Koudelková (Eds.), ICPE-EPEC 2013 Conference Proceedings. Active learning – In a changing world of new technologies (pp. 199–205). Charles University in Prague, MATFYZPRESS publisher.

Prediger, S. (2013). Darstellungen, Register und mentale Konstruktion von Bedeutungen und Beziehungen – mathematikspezifische sprachliche Herausforderungen identifizieren und bearbeiten. In M. Becker-Mrotzek, K. Schramm, E. Thürmann, & H. Vollmer (Eds.), Sprache im Fach – Sprachlichkeit und fachliches Lernen (pp. 167–183). Waxmann, Münster u.a.

Prediger, S., & Wessel, L. (2011). Darstellen – Deuten – Darstellungen vernetzen. Ein fach- und sprachintegrierter Förderansatz für mehrsprachige Lernende im Mathematikunterricht. In S. Prediger & E. Özdil (Eds.), Mathematiklernen unter Bedingungen der Mehrsprachigkeit – Stand und Perspektiven der Forschung und Entwicklung in Deutschland (pp. 163–184). Waxmann, Münster u. a.

Prediger, S., & Wessel, L. (2012). Darstellungen vernetzen. Ansatz zur integrierten Entwicklung von Konzepten und Sprachmitteln. Praxis der Mathematik in der Schule. Sekundarstufen I und II, 45 (Jg. 54), 28–33.

Rebello, S., & Cui, L. (2008). Retention and transfer of learning from mathematics to physics to engineering . American Society for Engineering Education.

Redish, E. F. (2016). Analysing the competency of mathematical modelling in physics. In G. Tomasz & E. Dbowska (Eds.), Key competences in physics teaching and learning (pp. 25–40). Cham: Springer International Publishing.

Redish, E. F., & Kuo, E. (2015). Language of physics, language of math: Disciplinary culture and dynamic epistemology. Science & Education, 24 (5–6), 561–590.

Schnotz, W., & Bannert, M. (2003). Construction and interference in learning from multiple representation. Learning and Instruction, 13 , 141–156.

Schulz von Thun, F. (n.y.). Das Kommunikationsquadrat . https://www.schulz-von-thun.de/die-modelle/das-kommunikationsquadrat (downloaded on 2018-06-11).

van Someren, M. (Hrsg.). (1998). Learning with multiple representations. Pergamon, as cited in Bauer (2015).

Strahl, A., Schleusner, U., Mohr, M., & Müller, R. (2010). Wie Schüler Formeln gliedern – eine explorative Studie. PhyDid, 1 (9), 18–24.

Tuminaro, J., & Redish, E. F. (2007). Elements of a cognitive model of physics problem solving: Epistemic games. Physical Review Special Topics-Physics Education Research, 3 (2), 020101.

Uhden, O. (2015). Verständnisprobleme von Schülerinnen und Schülern beim Verbinden von Physik und Mathematik. Zeitschrift für Didaktik der Naturwissenschaften, 22 , 1–12.

van der Meij, J., & Jong, T. d. (2006). Supporting students’ learning with multiple representations in a dynamic simulation-based learning environment. Learning and Instruction, 16 (3), 199–212. https://doi.org/10.1016/j.learninstruc.2006.03.007 , as cited in Bauer (2015).

Waldrip, B., Prain, V., & Carolan, J. (2010). Using multi-modal representations to improve learning in junior secondary science. Research in Science Education, 40 , 65–80.

Download references

Author information

Authors and affiliations.

Faculty of Physics, TU Dresden, Dresden, Germany

Marie-Annette Geyer & Wiebke Kuske-Janßen

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Marie-Annette Geyer .

Editor information

Editors and affiliations.

Technische Universität Dresden, Dresden, Germany

Gesche Pospiech

DCFA - Section of Mathematics and Physics, University of Udine, Udine, Italy

Marisa Michelini

The Weizmann Institute of Science Rehovot, Israel

Bat-Sheva Eylon

Rights and permissions

Reprints and permissions

Copyright information

© 2019 Springer Nature Switzerland AG

About this chapter

Cite this chapter.

Geyer, MA., Kuske-Janßen, W. (2019). Mathematical Representations in Physics Lessons. In: Pospiech, G., Michelini, M., Eylon, BS. (eds) Mathematics in Physics Education. Springer, Cham. https://doi.org/10.1007/978-3-030-04627-9_4

Download citation

DOI : https://doi.org/10.1007/978-3-030-04627-9_4

Published : 03 July 2019

Publisher Name : Springer, Cham

Print ISBN : 978-3-030-04626-2

Online ISBN : 978-3-030-04627-9

eBook Packages : Education Education (R0)

Share this chapter

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

ORIGINAL RESEARCH article

Investigating the concept of representation in the neural and psychological sciences.

A commentary has been posted on this article:

Commentary: Investigating the concept of representation in the neural and psychological sciences

• Read general commentary

• 1 Department of Philosophy, University of Central Florida, Orlando, FL, United States
• 2 Cognitive Sciences Program, University of Central Florida, Orlando, FL, United States
• 3 Department of History and Philosophy of Science, University of Pittsburgh, Pittsburgh, PA, United States
• 4 Center for Philosophy of Science, University of Pittsburgh, Pittsburgh, PA, United States
• 5 African Centre for Epistemology and Philosophy of Science, University of Johannesburg, Johannesburg, South Africa

The concept of representation is commonly treated as indispensable to research on brains, behavior, and cognition. Nevertheless, systematic evidence about the ways the concept is applied remains scarce. We present the results of an experiment aimed at elucidating what researchers mean by “representation.” Participants were an international group of psychologists, neuroscientists, and philosophers ( N  = 736). Applying elicitation methodology, participants responded to a survey with experimental scenarios aimed at invoking applications of “representation” and five other ways of describing how the brain responds to stimuli. While we find little disciplinary variation in the application of “representation” and other expressions (e.g., “about” and “carry information”), the results suggest that researchers exhibit uncertainty about what sorts of brain activity involve representations or not; they also prefer non-representational, causal characterizations of the brain’s response to stimuli. Potential consequences of these findings are explored, such as reforming or eliminating the concept of representation from use.

1. Introduction

The concept of representation is widely applied in research on brains, behavior, and cognition. We call this practice “mainstream representationalism.” Psychologists—especially cognitive psychologists—have historically investigated and explained mental capacities in terms of representations and the computations or operations that process them (e.g., Anderson, 1978 ; Chomsky, 1980 ; Fodor, 1981 ). Examples of this theoretical commitment abound in contemporary work as well, including research on attitudes (“attitudes … can be conceptualized as mental representations that determine how we evaluate stimuli”; De Houwer et al., 2021 , p. 870), concept learning (“[o]ur approach crucially exploits the classic insight that representational simplicity is a major determinant of learnability, with learners preferring to infer rules that are concise in their representational system”; Piantadosi et al., 2016 , p. 394; italics in original), and imagery (“[d]o learners who understand a picture also construct multiple mental representations in their mind”; Schnotz et al., 2021 , p. 4). The concept of representation is not only central to research on human mental capacities, but it is also widely applied in research on non-human animals (e.g., “[for] dogs, hearing an object’s verbal label evokes a mental representation of the object”; Dror et al., 2022 , p. 8) and on artificial intelligence architectures such as Adaptive Control of Thought-Rational (ACT-R; Anderson et al., 2004 ), Soar ( Laird, 2012 ), Semantic Pointer Architecture Unified Network (Spaun; Eliasmith et al., 2012 ), and deep neural networks (e.g., Barrett et al., 2019 ).

Similarly, the concept of representation plays a central role in the neural sciences. A central theoretical commitment is that brains form representations of the organism’s internal states (e.g., proprioceptive experiences), the external environment (e.g., speed and orientation of visual stimuli), or relational states that cross the internal/external dichotomy (perhaps rewards or beauty). Accordingly, neuroscientists commonly aim at identifying and characterizing these representations in order to answer questions such as the following: what do they represent, what are their vehicles, and how are they used (e.g., Kriegeskorte and Diedrichsen, 2019 ; Poldrack, 2021 )? This is especially true of subdisciplines such as the cognitive neurosciences (e.g., “[w]e usually take for granted the idea that information processing depends on internal representations”; Gazzaniga et al., 2014 , p. 74), computational neuroscience (e.g., “[c]learly, the brain must use specific representations and specific algorithms, and it is the goal of computational neuroscience to help find them”; Trappenberg, 2010 , p. 12), and sensory neurosciences (e.g., “[t]here is a complete representation of visual space in columns dominated by each eye”; Reid and Usrey, 2013 , p. 590).

Following suit, philosophers of psychology and neuroscience have proposed various explications of the concept of representation, sometimes inspired by traditional philosophy of mind (e.g., Von Eckardt, 1995 ; Ramsey, 2007 ; Egan, 2014 ; Shea, 2018 ), sometimes by work on signaling (e.g., Planer and Godfrey-Smith, 2021 ), and sometimes by the methods used by neuroscientists to identify neural representations, such as deep learning (e.g., Cao, 2022 ) and representational similarity analysis (e.g., Roskies, 2021 ). A minority—but increasingly vocal—group of psychologists (e.g., Richardson et al., 2008 ), neuroscientists (e.g., Buzsáki, 2019 ), and philosophers (e.g., Chemero, 2009 ; Hutto and Myin, 2013 ) disagree with this mainstream representationalism. They argue that the concept of representation need not be central, or even necessary, to investigate and explain brains, behavior, and cognition.

While the widespread appearance of the concept of representation in the neural and psychological sciences is indubitable, systematic evidence about the ways this concept is applied in these sciences remains scarce (for a rare exception; see Vilarroya, 2017 ). This article presents four preregistered studies that examine how researchers apply the concept of representation and five other ways of describing how the brain responds to stimuli. This project is in part descriptive: Our main goal was to examine empirically how the concept of representation is used in neural and psychological scientific practice. Additionally, projects such as the current one may also have normative implications. What is at stake is the theoretical status quo concerning the concept of representation, viz. , the widespread assumption in the neural and psychological sciences that the concept of representation is understood precisely enough to guide the development of hypotheses, interpretations of experimental data, and explanations. Systematically elucidating what researchers mean by “representation” may draw attention to imprecisions in the concept of representation, which could weaken the strength of conclusions drawn in research that hinges in crucial ways on its meaning. The imprecision of a scientific concept manifests itself in uncertainties concerning what follows from applying it (e.g., “What follows if some brain pattern is a representation?”) and what must be the case for this concept to apply (e.g., “What properties should a brain pattern have to count as a representation?”).

It is important to make clear that the abovementioned issues do not rest on the naive view that a concept can only be appropriately used in scientific research if it is defined by a widely accepted set of necessary and jointly sufficient conditions. Except for formal systems (e.g., logic and mathematics) or a handful of concepts (e.g., the concept of uncle), few concepts can be defined ( Machery, 2009 ), particularly concepts of entities and processes in the natural world. As such, there is no doubt that science progresses without defining all of its terms. Moreover, the absence of definitions can be viewed as an indispensable feature of research when scientists are attempting to characterize novel and interdisciplinary targets of the investigation, as has been the case in the investigations of genes and viruses (e.g., Rheinberger, 2000 ). Neurophilosopher Patricia Churchland captures well this idea when she writes that to “force precision by grinding out premature definitions enlightens nobody” ( Churchland, 1986 , p. 346).

While the imprecision of some uses of the concept of representation in the neural and psychological sciences can certainly be understood this way, such instances are not what we draw attention to. Consider the following two recent examples from neuroscience. First, in an article on finger movement, the concept of representation is used in contexts such as “different spatial representations,” “low-dimensional representation,” “n members can be represented at time t,” “schematic representation of behavioral mode segmentation,” “the cerebral cortex represents,” and “well-represented in neural state space” ( Flint et al., 2020 ). Second, an article on neural network models of symbolic cognitive processes and dynamical systems uses “representation” in contexts such as “agent’s internal representations of the environment,” “distributed representations,” “feature representation in deep learning,” “holographic reduced representations,” “neurobiological representations (i.e., grid cells),” and “structured symbolic representations” ( Voelker et al., 2021 ). One could reasonably be uncertain about what “representation” means in these instances and what would be required for something to be, for example, a “holographic reduced representation” or “represented in the neural state space.” How does a reader understand if it is reasonable to ask whether a structured symbolic representation ( Voelker et al., 2021 ) can be well-represented in neural state space ( Flint et al., 2020 )? Granted that these are merely two examples, our goal in this article was to show that they are illustrative of the kinds of imprecision commonly exhibited by uses of “representation” in the psychological and neural sciences. Note that our goal was not to bring to light disagreement about the meaning of “representation” or about variation in how scientists understand this expression across disciplines. On the contrary (as our results demonstrate below), there is little variation in the application of “representation” and related concepts across disciplines.

Moreover, such a project ought to be welcomed by both proponents and critics of mainstream representationalism. Proponents should welcome empirical descriptions of the uses of the concept of representation in order to regiment more tightly how it is used; critics should welcome these descriptions in order to develop more forceful critiques based on a better understanding of the roles the concept of representation plays in the theories of brain, behavior, and cognition that they have long argued does not require representations.

Some handpicked examples, such as those mentioned thus far, are insufficient to provide evidence about how neuroscientists or psychologists use “representation.” All the same, a literature review of the uses of “representation” in even one discipline would be a Herculean task. Thus, how can we empirically assess the current state of neuroscientists’ and psychologists’ understanding of the concept of representation? One option is to utilize what linguists call “elicitation studies” ( Greenbaum and Quirk, 1970 ). Instead of asking scientists to reflect and report on their own concepts or examining the natural occurrences of a given concept (e.g., corpus study), scientists are asked to use the target concept and then the experimenter can make inferences about its content based on subjects’ answers ( Machery, 2017 ; Machery et al., 2019 ). Inspired by the elicitation-study method, we conducted a survey-based experiment with an international group of neuroscientists, psychologists, and philosophers ( N  = 736). Since the number of respondents was too small to disaggregate the sample along participants’ subdisciplines, each group was analyzed as a combination of subdisciplines. For example, the group of neuroscientists included respondents who identified cellular and molecular neuroscience, cognitive neuroscience, and systems neuroscience as their subdisciplines. We tested the following four preregistered hypotheses:

1. Scientists will not demonstrate statistically significant differences among the concepts selected in the contrasts: representing, carrying information, being about, responding to, processing, and identifying.

2. Scientists will be sensitive to the specificity of content (i.e., high vs. low) but will not be sensitive to its functional integration or the nature of the vehicle (i.e., area vs. population of neurons).

3. On average, scientists will be willing to assign misrepresentations with their average responses falling within the range of either “strongly agree,” “agree,” or “somewhat agree.”

4. In the following three areas: (A) drawing distinctions between concepts; (B) being sensitive to the specificity and functional integration; and (C) being more willing to assign misrepresentation—philosophers will select for more distinctions (A), more sensitivity (B), and more willingness (C), than scientists as a group (i.e., cognitive scientists, psychologists, and neuroscientists).

The experiment consisted of four studies with the same basic structure. Participants were given a cover story about a neuroscientific study recording brain response to various stimuli, including faces and artifacts ( Figure 1 ). Participants were then asked to provide a rating on a 7-point scale (from 1: “Strongly agree” to 7: “Strongly disagree”) regarding six questions about whether they would agree to describe the brain’s activity as representing , carrying information , being about , responding , processing , and identifying the stimuli.

Figure 1 . Sample experimental stimuli. The experiment consisted in four studies of similar design, each with a cover story like the one associated with this figure: “In a study published about ten years ago, participants were presented with visual stimuli in a standard block design with alternating images of human faces and houses (A) . Data were obtained via a microelectrode (B) from single neurons in participants’ fusiform face area (C) . An example of the time series data obtained during the task is presented in (D) .” Modified and reprinted with permission from Michael J. Tarr, PxHere. CC0 1.0, flickr. CC BY 2.0 (A) ; ( Jia et al., 2013 ). CC BY 2.0 (B) ; Public domain (C) ; ( Alkan et al., 2011 ). CC BY 4.0 (D) .

These six terms were selected in order to provide participants with ways of describing the brain’s response to stimuli that (1) are used in neuroscience and psychology and (2) are of three different kinds: They describe this response as having an “intentional” component (“representing,” “being about,” and “identifying”), in causal terms (“responding” and “processing”), or as having an information-theoretic nature (“carrying information”). Our goal was to examine whether ways of describing the brain’s response to stimuli that are similar are treated similarly by scientists and to compare scientists’ willingness to use different kinds of descriptions (see Hypothesis 1).

The first term, “representing,” expresses the concept we are investigating. The examples from psychology and neuroscience given above are but a few of the countless instances of researchers using the term in peer-reviewed work. The third term, “about,” or the clause “being about,” describes the brain’s response as intentional. In the current context, being “intentional” means, roughly, standing for something else in the way a sign or a symbol stands for something else. Both describing a response to stimuli as “representing” or as “being about” something assume that this response is correctly understood as intentional (e.g., Baker et al., 2022 , p. 945). Aboutness is a central topic in the philosophy of mind (e.g., Yablo, 2014 ). Responses to stimuli are often described as being about something in psychology, for instance, concerning topics such as language acquisition ( Hurford, 2007 , p. 173: “[t]he aboutness, or Intentionality, of modern human utterances, derives from the aboutness or Intentionality of pre-linguistic mental representations”) and memory ( Klein, 2015 , p. 2: “coming from the past does not sanction the inference that the ‘something’ in awareness is about the past”), although the expression is less frequent in neuroscience (except in the context of “information about”). The sixth term, “identifying,” refers to an intentional activity: To identify something (e.g., a face) is more than just responding to some stimuli causally; it involves representing stimuli as something (e.g., as faces). It is sometimes used to describe what the brain does. For instance, Tarr and Gauthier (2000 , p. 764) began their influential literature review of the fusiform face area (FFA) by asking “How does the primate visual system process and identify objects?”

The fourth and fifth terms, “responding” and “processing,” differ from the three terms just discussed in describing the brain’s response to stimuli causally, without necessarily implying that this response has an intentional nature. “Responding” is often used in neuroscience, for instance, in research on visual perception (e.g., Barlow et al., 1964 ; Tarr and Gauthier, 2000 , p. 765), with an illustrative phrase being, “retinotopically organized map of neurons responding to parts of visual space” ( Ballard, 2015 , p. 144). The fifth term, “processing,” is also common (e.g., Tarr and Gauthier, 2000 , p. 768: “involved in processing subordinate-level information for all objects, including faces”).

Finally, we use the second term, “carrying information,” to examine whether it would behave more like intentional terms such as “representing,” more like causal terms like “responding,” or in a sui generis manner, perhaps aligned with Shannon and Weaver’s information theory ( Shannon and Weaver, 1949/1964 ). Information-theoretic descriptions of brains’ response to stimuli are very common in neuroscience, including computational neuroscience (e.g., Soh et al., 2018 ) and cognitive neuroscience (e.g., Ghuman et al., 2014 , p. 2: “Recent studies have demonstrated that the FFA activity contains information about individual faces”; Piazza and Eger, 2016 , p. 268: “the precise kind of number-related information that is encoded in that part of the brain”).

While these six terms are not exhaustive of all potentially relevant concepts we could have used, for the purposes of our experimental design they capture diverse ways of thinking about the brain’s response to stimuli across the neural and psychological sciences.

The goal of Study 1 was to examine whether neuroscientists, psychologists, and philosophers make any assumptions about the scale at which the vehicles of neural representations, that is, the brain substrates that represent stimuli, are to be found. Participants were randomly assigned to one of two conditions. In the neuron condition, they were told that the response of a neuron was measured by means of a microelectrode (visually represented) when presented with faces; in the population condition, they were told that the response of a neural population (what we refer to in the current study as an “area” 1 ; i.e., the fusiform face area (FFA)) was measured by means of functional magnetic resonance imaging (fMRI).

The goal of Study 2 was to examine what kind of relation, if any, must hold between the brain and stimuli for neuroscientists, psychologists, and philosophers to describe it in various terms. Participants were randomly assigned to one of two conditions. In the high specificity condition, a brain area, whose activity was measured by means of fMRI, responded to faces and only to them (it is perfectly specific or selective); in the low specificity condition, it responded to faces but also to houses.

The goal of Study 3 was to examine whether evidence that the brain’s response to stimuli is used by a broader neural network and thus has a function ( Cummins, 1975 ), in addition to a perfect correlation with a stimulus increases neuroscientists, psychologists, and philosophers’ willingness to treat the brain’s response to stimuli in representational terms. It is common to distinguish two kinds of function: teleological vs. “Cummins-style” function (e.g., Millikan, 1989 ). The teleological function of an object (e.g., an organ or a component of an artifact) explains why this object exists by identifying what it does. In this sense, the function of the heart is to pump blood in the body and the function of glasses is to focus light at the right place of the eye’s lens. Philosophers of biology have often appealed to natural selection and other selective processes (e.g., culture or development) to explain how an object could have a function in this sense (e.g., Millikan, 1989 ). The Cummins-style function of a part of a system describes how this part causally influences the broader system it is a part of. Study 3 focuses on this notion of function: By embedding the brain area in a broader neural network, our goal was to suggest that the brain causally contributes to a broader system involved in face recognition and, thus, make it clear that it has a particular Cummins-style function. In the mere correlation condition, participants were just given evidence of the brain’s response to the stimuli; in the function condition, the connection between the relevant brain area and a full network was highlighted verbally and by means of two figures.

Finally, the goal of Study 4 was to examine whether neuroscientists, psychologists, and philosophers are willing to describe the brain’s response as erroneous, for example, whether it misrepresents stimuli. Philosophers concur that for a state to count as a representation, misrepresentation must be possible (e.g., Ramsey, 2007 , p. 12; Shea, 2018 , p. 10). Participants were assigned to a single condition where a brain area that responds to faces happens to also respond, once, to a house.

These four studies focus on characteristics that brain states would have to possess if they are to count as representations. Representations must occur at some scale in brain organization (Study 1); the occurrence of representations must causally depend, in some way, on what they represent (Study 2); representations must be used by downstream processes (Study 3); and representations can be misapplied (Study 4). To have a precise concept of representation is to have a sense of the scale at which representations occur, of the nature of representations’ causal dependence on what they represent, and on the significance of the use of representations, or at least to have some sense for some of these issues. Additionally, a precise concept of representation ought to distinguish cases for which misapplication matters and those for which it does not (Study 4). In what follows, we report the results from each study.

2.1. Participants

The study was approved by the Institutional Review Boards at the University of Central Florida (IRB STUDY00002612) and the University of Pittsburgh (IRB STUDY20050065). All research was performed in accordance with relevant guidelines/regulations. Hypotheses and data collection methods including the stopping rule, exclusion criteria, and data analytic strategies were preregistered with the Open Science Framework (OSF; https://osf.io/mskwy/ ; doi: 10.17605/OSF.IO/SARVU).

Two research assistants were tasked to create a database of emails found on the public websites of departments, centers, institutes, and schools at universities around the world. A list of universities in Asia, Australia, Europe, North America, and South America was created, and the research assistants were asked to input the names, emails, and departmental affiliations of cognitive scientists, computer scientists, linguists, neuroscientists, philosophers, and psychologists into a data file. Research assistants were ultimately asked to focus on cognitive scientists, neuroscientists, and psychologists in the United States, setting aside computer scientists and linguists as well as academics from abroad. A total of 14,338 recruitment emails were sent, many of which were blocked by university servers. As was indicated in the preregistration, the study was also advertised on blogs, mailing lists, and social media.

In total, 736 participants completed the study. We excluded participants who reported being younger than 18, who were not graduate students, postdoctoral researchers, professors, or researchers with a doctorate, who either did not respond or gave an incorrect answer to the last question of the survey, “Please tell us what this study was about,” and who provided the same numerical answer to questions in all four scenarios (in line with the preregistration). We also limited our analysis to neuroscientists, psychologists, and philosophers ( Table 1 ), setting aside cognitive scientists in light of the small number of participants who self-identified as such and completed the study (52 before exclusion; a departure from the preregistration).

Table 1 . Demographic characteristics of neuroscientists, psychologists, and philosophers.

2.2. Materials

The recruitment materials included a link to a survey on Qualtrics. Participants were first asked a few demographic questions before being asked to complete successively four studies in a random order (described below). They were then asked several philosophical questions related to representation, computation, and their broader commitments related to the foundations of neuroscience and cognitive science [full survey available at the preregistration site ( https://osf.io/mskwy/ ; doi:10.17605/OSF.IO/SARVU)].

Each of the four studies had the same basic structure. Participants were given a cover story about a neuroscientific study measuring brain response to various stimuli, including faces and artifacts. The first figure represented the basic structure of the experimental design. Additional figures represented the data observed, including a time series. Participants were then asked six questions about whether they would agree to describe the brain’s activity as representing the stimuli, carrying information about the stimuli, being about the stimuli, responding to the stimuli, processing the stimuli, and identifying the stimuli (each on a 7-point scale anchored at “1” with “strongly agree”).

2.3. Availability of data and materials, and analyses

The datasets generated and/or analyzed during the current study are available in the Open Science Framework (OSF) repository ( https://osf.io/mskwy/ ; doi:10.17605/OSF.IO/SARVU). All analyses were conducted on R (script available at the preregistration site: https://osf.io/mskwy/ ; doi: 10.17605/OSF.IO/SARVU). As preregistered, the significance level was set at 0.005 ( Benjamin et al., 2018 ). p -values between 0.05 and 0.005 are taken to be suggestive and in need of confirmation. All the analyses were redone with participants who had completed a PhD. The results did not change.

3.1. Mainstream representationalism

Toward the end of the survey, participants were asked five questions aimed at elucidating positions on foundational issues concerning the nature of cognition. We begin by reporting the results from a question probing their commitment to mainstream representationalism: “Does cognition involve representations? Yes or no.” We claimed at the start that mainstream representationalism, i.e., mental capacities involve computations acting on representations and that brains represent stimuli—is widely accepted as being necessary to investigate and explain brains, behavior, and cognition. As expected, a very large majority of participants answered this question positively for the three disciplines of interest ( Figure 2 ). It thus appears that mainstream representationalism is embraced by a large majority of psychologists, neuroscientists, and philosophers.

Figure 2 . The proportion of “No” and “Yes” answers to the representation question. The overwhelming majority of neuroscientists, philosophers, and psychologists answered “yes” to the question, “Does cognition involve representations?”

3.2. Study 1: vehicles of representations

The distribution of responses is presented in Figure 3A . A mixed-design analysis of variance (ANOVA) with questions as a within-participant factor (six levels), discipline as a between-participant factor (three levels), and condition as a between-participant factor (two levels) revealed a main effect ( Benjamin et al., 2018 ) of question ( F (5, 3,083) = 167.8, p  < 0.001, η 2  = 0.2), a suggestive effect of discipline ( F (2, 3,083) = 5.0, p  = 0.007, η 2  = 0.003), and no effect of condition ( F (1, 3,083) = 3.5, p  = 0.06, η 2  ≤ 0.001). Post-hoc analysis revealed that the suggestive effect observed for discipline is due to a suggestive difference between philosophers and psychologists ( t (3083) = 3.0, p  = 0.007, d  = 0.14); no other comparison reaches the 0.05 level. All post-hoc comparisons between the six questions used were significant except for the non-significant comparison between represents and identifies ( t (3083) = 0.9, p  = 0.9) and for the suggestive comparison between is about and identifies ( t (3083) = −3.3, p  = 0.01). The main effects of discipline and question were qualified by a suggestive two-way interaction ( F (10, 3,083) = 2.4, p  = 0.007, η 2  = 0.006; Figure 3B ).

Figure 3 . Study 1: Vehicles of representations. Distribution of answers for Study 1 (1: “Strongly agree;” 7: “Strongly disagree”) (A) . Interaction of question and discipline in Study 1 (B) . Error bars correspond to 95% confidence intervals.

In line with this interaction, an exploratory, not-preregistered mixed-design ANOVA with the question as a within-participant factor (six levels) and condition as a between-participant factor (two levels), was conducted for neuroscientists and psychologists separately. For neuroscientists, we observed a main effect of question ( F (5, 1,049) = 63.5, p  < 0.001, η 2  = 0.23), no effect of condition, and no interaction (both p s > 0.7). All post-hoc comparisons were significant except for the non-significant comparisons between represents and is about, represents and identifies, and is about and processes ( p s > 0.5) and for the suggestive comparisons between represents and processes and is about and identifies ( p s > 0.01). For psychologists, we observed a main effect of question ( F (5, 941) = 53.9, p  < 0.001, η 2  = 0.22), no effect of condition, and no interaction (both p s > 0.8). All post-hoc comparisons were significant except for the non-significant comparisons between represents and identifies, is about and processes, carries information and processes (all p s > 0.15), is about and carries information, and is about and identifies ( p  = 0.055 and 0.057, respectively).

Three main findings emerge from this first study. First, contrary to our first preregistered hypothesis, neuroscientists and psychologists do not treat all of the descriptions of the brain’s response to stimuli identically. The results indicate that neuroscientists and psychologists find acceptable lean, causal characterizations of the brain’s response to stimuli in terms of responding and processing, as well as an information-theoretic characterization (carrying information about). By contrast, they appear uncertain about intentional characterizations. On average, participants chose “neither agree nor disagree” for “representing,” “identifying,” and “being about.” We will come back to this point in the general discussion below. Importantly, neuroscientists’ and psychologists’ overall uncertainty is not the result of a bimodal distribution, which would be indicated by half of the participants willing to strongly agree to use the concept of representation to describe the brain’s response to stimuli and half of them strongly disagreeing. Rather, the distribution is centered around its mean (The same is true of the three other studies.).

Second, the results suggest that it made very little difference to neuroscientists and psychologists whether the vehicle of representation was verbally and pictorially represented as a single neuron or as a brain area. This negative result suggests that neuroscientists and psychologists do not have any expectations about the scale at which representations are to be found in the brain. In other words, they may subscribe to mainstream representationalism, but their concept of representation is not specific enough to dictate what kind of brain structure or pattern at what level of aggregation (neuron, population, and distributed network of populations) would be a representation.

Third, while philosophers were somewhat less likely to agree with our prompts than psychologists, the variation across disciplines was small. This finding suggests that the concept of representation has not specialized in the disciplines we are considering (see Machery et al., 2019 for a discussion of similar results for the concept of innateness in psychology, biology, and linguistics).

3.3. Study 2: specificity and representation

The distribution of responses is presented in Figure 4A . A mixed-design ANOVA with question as a within-participant factor (six levels), discipline as a between-participant factor (three levels), and condition as a between-participant factor (two levels) revealed a main effect of question ( F (5, 3,083) = 191.4, p  < 0.001, η 2  = 0.23), a suggestive effect of discipline ( F (2, 3,083) = 4.9, p  = 0.007, η 2  = 0.002), and an effect of condition ( F (1, 3,083) = 131.3, p  < 0.001, η 2  = 0.03). Post-hoc analysis revealed that the suggestive effect observed for discipline is due to a difference between neuroscientists and philosophers ( t (3083) = 2.6, p  = 0.03, d  = 0.1) and neuroscientists and psychologists ( t (3083) = 3.0, p  = 0.008, d  = 0.14). All post-hoc comparisons between the six questions used were significant except for the non-significant comparisons between represents and identifies ( t (3083) = 2.2, p  = 0.2), between carries information and processes ( t (3083) = −1.0, p  = 0.9), and between is about and identifies ( t (3083) = −0.7, p  = 0.98) and for the suggestive comparison between represents and is about ( t (3083) = 3.1, p  = 0.03). The main effects of discipline and condition were qualified by a two-way interaction ( F (10, 3,083) = 9.0, p  < 0.001, η 2  = 0.004): Psychologists are more sensitive to the manipulation of specificity than philosophers and neuroscientists.

Figure 4 . Study 2: Specificity and representation. Distribution of answers for Study 2 (1: “Strongly agree”; 7: “Strongly disagree”) (A) . Interaction of condition and discipline in Study 2 (B) . Error bars correspond to 95% confidence intervals.

In addition, we explored the impact of specificity on representation alone ( Figure 4B ). For neuroscientists, the impact of specificity on the description of the brain’s response in terms of representation was too small to result in a significant or suggestive effect ( t (173.07) = −1.90; p  = 0.059, d  = 0.1); by contrast, we found a significant effect for psychologists ( t (155.12) = −5.7; p  < 0.001, d  = 0.3).

Two main findings emerge from Study 2. First, as we observed in Study 1, the results indicate preferences by neuroscientists and psychologists for thin, causal descriptions of the brain’s response to stimuli (responds to and processes) and information-theoretic descriptions over intentional descriptions and uncertainty about the latter. Second, specificity matters in describing how the brain responds to stimuli (in line with the preregistered second hypothesis). When one aggregates across ways of describing the brain’s response, neuroscientists, psychologists, and philosophers agree more (although to a different degree) when the brain’s response is maximally sensitive. Turning to the concept of representation, we only found evidence for the significance of specificity for psychologists. It would, thus, seem that psychologists take specificity to be relevant to whether some brain state can count as a representation. However, even perfect specificity does not appear to lead psychologists to express certainty when it comes to describing the brain’s response to stimuli in representational or, more generally, intentional terms.

3.4. Study 3: function and representation

The distribution of responses is presented in Figure 5 . A mixed-design ANOVA with question as a within-participant factor (six levels), discipline as a between-participant factor (three levels), and condition as a between-participant factor (two levels) revealed a main effect of question ( F (5, 3,083) = 150.2, p  < 0.001, η 2  = 0.19), and suggestive effects of discipline ( F (2, 3,083) = 4.7, p  = 0.009, η 2  = 0.002) and condition ( F (1, 3,083) = 6.9, p  = 0.009, η 2  = 0.002), but no interaction. Post-hoc analysis revealed that the suggestive effect observed for discipline is due to a suggestive difference between neuroscientists and psychologists ( t (3083) = 2.9, p  = 0.009, d  = 0.1); no other comparison was significant at the 0.05 level. All post-hoc comparisons between the six questions used were significant except for the non-significant comparisons between carries information and processes ( t (3083) = 2.0, p  = 0.4) and between is about and identifies ( t (3083) = −1.0, p  = 0.9) and for the suggestive comparison between represents and identifies ( t (3083) = 3.2, p  = 0.02). To explore the role of function in the assignment of representation, we conducted an ANOVA with question as a within-participant factor (six levels) and condition as a between-participant factor (two levels). No significant or suggestive effect was observed.

Figure 5 . Study 3: Function and representation. Distribution of answers for Study 3 (1: “Strongly agree”; 7: “Strongly disagree”).

Two main findings emerge from Study 3. First, as was found in Studies 1 and 2, the results indicate preferences by neuroscientists and psychologists for thin, causal vocabulary to describe the brain’s response to stimuli and uncertainty about intentional vocabulary. Second, whether or not the brain area’s response to a stimulus is embedded in a larger network, and thus, whether it has a function, influenced how the brain’s response was described, although it did not appear to influence whether it was described in representational terms. When it comes to representation, we found no evidence that having a function matters (in line with the preregistered second hypothesis).

3.5. Study 4: misrepresentation

The distribution of responses is presented in Figure 6 . A mixed-design ANOVA with question as a within-participant factor (six levels) and discipline as a between-participant factor (three levels) revealed a main effect of question ( F (5, 3,101) = 11.5, p  < 0.001, η 2  = 0.02) and discipline ( F (2, 3,101) = 26.6, p  < 0.001, η 2  = 0.02), but no interaction. Post-hoc analysis revealed that the effect observed for discipline is due to significant differences between all disciplines, with philosophers being less unwilling to view the brain’s response as erroneous (in line with the fourth preregistered hypothesis). All pairwise comparisons between questions were significant at the 0.005 level, except for represents and is about, represents and identifies, carries information and represents, carries information and processes, is about and identifies, and responds and processes, which were not significant at the 0.05 level, and carries information and identifies and processes and identifies, which were only suggestive (0.01 <  p s < 0.05). We also found that neuroscientists and psychologists are unwilling to assign misrepresentation (both mean answers significantly higher than the neutral point, “neither agree nor disagree”; p s < 0.001).

Figure 6 . Study 4: Misrepresentation. Distribution of answers for Study 4 (1: “Strongly agree”; 7: “Strongly disagree”).

The main finding to emerge from Study 4 is that neuroscientists, and to a smaller extent psychologists, did not describe the brain’s response to stimuli as erroneous, that is, as failing to do what the brain is meant to do (contrary to the preregistered third hypothesis). In particular, neuroscientists and psychologists appear unwilling to say that it misrepresents something as something else.

4. Discussion

Neuroscientists, psychologists, and philosophers commonly embrace the idea that the concept of representation is indispensable to investigating and understanding brains, behavior, and cognition. While neuroscientists, psychologists, and philosophers occasionally differ in their responses to the experimental stimuli used in our four studies, those differences appear to be very small. The concept of representation does not appear to have specialized in different directions as scientific concepts sometimes do when they are used in different disciplines ( Hull, 1988 ; Machery et al., 2019 ). While our sample size was not large enough to investigate whether the concept of representation varies within disciplines (e.g., between molecular and system neuroscientists), we found no evidence for this since the data were not bimodally distributed. However, further work should address this limitation of the present study and compare neuroscientists across subdisciplines.

Furthermore, having a precise concept of representation requires having some sense of what follows from something being a representation (or of what is required for something to count as a representation), including the scale at which it occurs, the way it depends on stimuli, or how it features in downstream processes. Additionally, having a precise concept of representation would facilitate distinguishing representations from other kinds of signs. Despite the centrality of representations for investigating and explaining brains, cognition, and behavior ( Figure 2 ), findings from Studies 1 to 4 suggest that “representation” may express an imprecise concept among researchers.

First, in none of the four studies did neuroscientists and psychologists describe the brain’s response as representing its stimulus. One interpretation of this finding is that participants do not do so because they think this response is not an instance of representation. However, another reason may be their uncertainty about what is required for something to be a representation as illustrated by their neither agreeing nor disagreeing. This pattern is found in other intentional descriptions such as the idea that the brain’s response is about its stimulus or what it identifies as its stimulus. This uncertainty stands in contrast with neuroscientists’ and psychologists’ selections of thinner, causal descriptions, such as responding and processing, to the brain’s response to stimuli. Neuroscientists and psychologists also selected descriptions of the brain’s response in information-theoretic terms, suggesting perhaps that they understand information in more a causal sense than an intentional one.

Second, neuroscientists and psychologists do not appear to have a precise idea about what kind of brain structure or pattern counts as representation. Whether the brain’s response was described at the neuronal (single neuron) or at the population level (what we refer to in the current study as an “area”) made little difference to their answers.

Third, neuroscientists appear not to require the brain’s response to be used in a broader neural network and, thus, to have a function ( Cummins, 1975 ) to count as a representation. They could be indifferent to the role of function for representations either because they endorse a non-functionalist, correlation-based account of representation or because they are uncertain about what is required for something to count as a representation. Their uncertainty in applying the concept of representation noted above suggests that the latter is more likely the case. For psychologists, on the other hand, representation requires specificity, that is, brain states cannot be representations if they occur in response to different types of stimuli. Thus, psychologists’ concept of representation is more precise than neuroscientists’ concept: They appear to endorse a necessary condition for the application of this concept.

These first three points tentatively suggest that psychologists’ and, to an even greater extent, neuroscientists’ concept of representation is imprecise: Psychologists and neuroscientists selected responses indicating uncertainty about what properties a brain pattern must have to count as a representation and what follows from calling a brain pattern a representation. This uncertainty extends to other intentional notions and contrasts with thinner, causal notions.

One of the few things philosophers working on representation agree upon is that representation requires misrepresentation (e.g., Bechtel, 1998 ; Haugeland, 1998 ; Ramsey, 2007 ; Shea, 2018 ), that is to say, representations can be misapplied; for example, a map can misrepresent the region it is about; we can call a dog a “wolf.” By contrast, a natural sign cannot misrepresent ( Dretske, 1988 ): The smoke produced by the fire carries information about the fire, but it cannot misrepresent it; tree rings carry information about the age of the tree but cannot misrepresent it; and so on. Neuroscientists and psychologists did not select responses that describe the brain’s response as erroneous, including as being a misrepresentation. These choices suggest that their concept of representation may be imprecise to a degree that does not distinguish natural signs and representations.

If, as our results suggest, there is indeed widespread uncertainty in applications of the concept of representation, then such a state of affairs might not be innocuous. They could breed fruitless debates about whether or not some brain part that responds to some stimulus represents it; barring a clearer concept of representation, such debates cannot be resolved. For instance, in the embodied cognition literature, cognitive neuroscientists have provided ample fMRI evidence that at least sometimes (e.g., Kiefer and Pulvermüller, 2012 ), motor and perceptual areas of the brain are activated when participants retrieve and use concepts, but critics have responded that those activations are incidental: They are not the conceptual representations themselves (e.g., Mahon and Caramazza, 2008 ). Without greater precision about what it means for a brain pattern to be a representation and some operationalization of the concept of representation, this controversy is unlikely to be resolved. Furthermore, imprecision of the concept of representation could prevent neuroscientists from interpreting some experimental results univocally. fMRI adaptation, multi-voxel pattern analysis (MVPA), representational similarity analysis, and others are supposed to determine what kind of representations the brain produces and where. If the concept of representation at play is genuinely imprecise, then it is hard to say what such methods reveal about the brain.

What is to be done with an imprecise scientific concept such as, possibly, the concept of representation? One approach is that such concepts must be reformed or, as philosophers say, “explicated” ( Carnap, 1950 ), “prescriptively analyzed” ( Machery, 2017 ), or “engineered” ( Cappelen, 2018 ). Explication takes an existing concept (either a folk or a scientific concept) and improves it, often in order to use it in philosophical or scientific theorizing. Another approach, well-known in the history and philosophy of science, is to propose to eliminate the concept of representation from neuroscience and psychology. In the current context, concepts are eliminated from scientific theorizing when that concept does not refer to anything that actually exists (e.g., the concept of phlogiston in a theory of combustion; Churchland and Churchland, 1998 ) or enables discourse that is misleading or problematic (e.g., perhaps the concept of qualia in a theory of consciousness; Dennett, 1993 ; for more on elimination, see, e.g., Churchland (1979) for folk psychological concepts; Griffiths (1997) for the concept of emotion; Griffiths et al. (2009) for the concept of innateness).

Given the extensive use of the concept of representation, it is reasonable to conclude that most neuroscientists and psychologists would strongly prefer the former option, and it is likely that most philosophers of psychology and neuroscience would agree. At the very least, elimination might be impracticable and, at most, quite costly. Still, one might push for the elimination of the concept of representation, an option critics of mainstream representationalism in psychology and neuroscience would prefer. If the concept of representation is to be eliminated, neuroscience would have to put its results, methods, and theories in non-representational terms. While the exact shape of future neuroscience cannot be predicted, it is worth noting that, following our findings that neuroscientists are willing to describe the brain’s response in causal and informational terms, the tools already exist to describe the dynamics of neural processes in non-representational terms (e.g., Izhikevich, 2007 ; Honey and Sporns, 2008 ; Shenoy et al., 2013 ; Sussillo and Barak, 2013 ; Cunningham and Byron, 2014 ; Dumas et al., 2014 ; Zhang et al., 2020 ; for additional review see Favela, 2020 ; Favela, 2021 ).

If reforming is the better option, then doing so for the concept of representation would require specifying to a sufficient degree of precision the characteristics of representation that make something a representation, including its use and its causal dependence on what it represents, and it would distinguish representations from natural signs. Similar to the above discussion regarding the point that both proponents and critics of mainstream representationalism should welcome the current set of studies—albeit, for different purposes—proponents ought to welcome opportunities for widespread reforming of the concept of representation. For example, identifying ways various usages of the concept are misleading or problematic could facilitate a more certain and precise discourse, which would, in turn, enable more fruitful research. While we remain neutral here about which of these two options is preferable, the current study lends support to the idea that the concept of representation requires precisification, work that will benefit our collaborative interests in understanding brains, behavior, and cognition.

Data availability statement

The datasets analyzed for this study can be found in the Open Science Framework (OSF; https://osf.io/mskwy/ ; doi: 10.17605/OSF.IO/SARVU).

Ethics statement

The studies involving human participants were reviewed and approved by Institutional Review Boards at the University of Central Florida (IRB STUDY00002612) and the University of Pittsburgh (IRB STUDY20050065). No consent was required for this study by the Institutional Review Boards.

Author contributions

LF and EM contributed equally to the manuscript. All authors contributed to the article and approved the submitted version.

Acknowledgments

The authors thank our research assistants Maya Best and Dean Allen Walters for their support with data collection. For providing helpful feedback on earlier versions of the experiment’s survey and manuscript, the authors thank John Beggs, John Bickle, Tony Chemero, Stephen Fiore, and John Krakauer, the X-Phi Lab at the University of Pittsburgh, and the Fellows reading group at the Center for Philosophy of Science (Leonardo Bich, Ravit Doran, Heather Douglas, Eugen Fischer, Ruth Kastner, Laura Menatti, Aydin Mohseni, and Serife Tekin). The authors thank Mary Jean Amon for discussions concerning data analyses and assistance with figures. The authors also thank audiences for feedback on the project from the Philosophy and Neuroscience at the Gulf IV: Fourth Annual Meeting of the Deep South Philosophy and Neuroscience Workgroup and the 58th Annual Meeting of the Alabama Philosophical Society, Neural Mechanisms Online 2022, and the 3rd Joint Conference of the Society for Philosophy and Psychology and European Society for Philosophy and Psychology.

Conflict of interest

The authors declare that the study was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

1. ^ As a reviewer usefully pointed out, this terminology is a simplification: One can identify brain areas by non-neuronal means.

Alkan, Y., Biswal, B. B., and Alvarez, T. L. (2011). Differentiation between vergence and saccadic functional activity within the human frontal eye fields and midbrain revealed through fMRI. PLoS One 6:e25866. doi: 10.1371/journal.pone.0025866

PubMed Abstract | CrossRef Full Text | Google Scholar

Anderson, J. R. (1978). Arguments concerning representations for mental imagery. Psychol. Rev. 85, 249–277. doi: 10.1037/0033-295X.85.4.249

CrossRef Full Text | Google Scholar

Anderson, J. R., Bothell, D., Byrne, M. D., Douglass, S., Lebiere, C., and Qin, Y. (2004). An integrated theory of the mind. Psychol. Rev. 111, 1036–1060. doi: 10.1037/0033-295X.111.4.1036

Baker, B., Lansdell, B., and Kording, K. P. (2022). Three aspects of representation in neuroscience. Trends Cogn. Sci. 26, 942–958. doi: 10.1016/j.tics.2022.08.014

Ballard, D. H. (2015). Brain Computation as Hierarchical Abstraction . Cambridge, MA: The MIT Press.

Google Scholar

Barlow, H., Hill, R. M., and Levick, W. R. (1964). Retinal ganglion cells responding selectively to direction and speed of image motion in the rabbit. J. Physiol. 173, 377–407. doi: 10.1113/jphysiol.1964.sp007463

Barrett, D. G., Morcos, A. S., and Macke, J. H. (2019). Analyzing biological and artificial neural networks: challenges with opportunities for synergy? Curr. Opin. Neurobiol. 55, 55–64. doi: 10.1016/j.conb.2019.01.007

Bechtel, W. (1998). Representations and cognitive explanations: assessing the dynamicist’s challenge in cognitive science. Cogn. Sci. 22, 295–318. doi: 10.1207/s15516709cog2203_2

Benjamin, D. J., Berger, J. O., Johannesson, M., Nosek, B. A., Wagenmakers, E. J., Berk, R., et al. (2018). Redefine statistical significance. Nat. Hum. Behav. 2, 6–10. doi: 10.1038/s41562-017-0189-z

Buzsáki, G. (2019). The Brain From Inside Out . New York, NY: Oxford University Press.

Cao, R. (2022). Putting representations to use. Synthese 200, 1–24. doi: 10.1007/s11229-022-03522-3

Cappelen, H. (2018). Fixing Language: An Essay on Conceptual Engineering . Oxford, UK: Oxford University Press.

Carnap, R. (1950). Logical Foundations of Probability . Chicago, IL: University of Chicago Press.

Chemero, A. (2009). Radical Embodied Cognitive Science . Cambridge, MA: MIT Press.

Chomsky, N. (1980). Rules and representations. Behav. Brain Sci. 3, 1–15. doi: 10.1017/S0140525X00001515

Churchland, P. M. (1979). Scientific Realism and the Plasticity of Mind . Cambridge: Cambridge University Press.

Churchland, P. S. (1986). Neurophilosophy: Toward a Unified Science of the Mind/Brain . Cambridge, MA: MIT Press.

Churchland, P. M., and Churchland, P. S. (1998). On the Contrary: Critical Essays, 1987–1997 . Cambridge, MA: The MIT Press.

Cummins, R. (1975). Functional analysis. J. Philos. 72, 741–765. doi: 10.2307/2024640

Cunningham, J. P., and Byron, M. Y. (2014). Dimensionality reduction for large-scale neural recordings. Nat. Neurosci. 17, 1500–1509. doi: 10.1038/nn.3776

De Houwer, J., Van Dessel, P., and Moran, T. (2021). Attitudes as propositional representations. Trends Cogn. Sci. 25, 870–882. doi: 10.1016/j.tics.2021.07.003

Dennett, D. C. (1993). “Quining qualia” in Readings in Philosophy and Cognitive Science . ed. A. I. Goldman (Cambridge, MA: The MIT Press), 381–414.

Dretske, F. (1988). Explaining Behavior: Reasons in a World of Causes . Cambridge, MA: The MIT Press.

Dror, S., Sommese, A., Miklósi, Á., Temesi, A., and Fugazza, C. (2022). Multisensory mental representation of objects in typical and gifted word learner dogs. Anim. Cogn. 25, 1557–1566. doi: 10.1007/s10071-022-01639-z

Dumas, G., de Guzman, G. C., Tognoli, E., and Kelso, J. S. (2014). The human dynamic clamp as a paradigm for social interaction. Proc. Natl. Acad. Sci. 111, E3726–E3734. doi: 10.1073/pnas.1407486111

Egan, F. (2014). How to think about mental content. Philos. Stud. 170, 115–135. doi: 10.1007/s11098-013-0172-0

Eliasmith, C., Stewart, T. C., Choo, X., Bekolay, T., DeWolf, T., Tang, Y., et al. (2012). A large-scale model of the functioning brain. Science 338, 1202–1205. doi: 10.1126/science.1225266

Favela, L. H. (2020). Dynamical systems theory in cognitive science and neuroscience. Phil. Compass 15, 1–16. doi: 10.1111/phc3.12695

Favela, L. H. (2021). The dynamical renaissance in neuroscience. Synthese 199, 2103–2127. doi: 10.1007/s11229-020-02874-y

Flint, R. D., Tate, M. C., Li, K., Templer, J. W., Rosenow, J. M., Pandarinath, C., et al. (2020). The representation of finger movement and force in human motor and premotor cortices. eNeuro 7, ENEURO.0063–ENEU20.2020. doi: 10.1523/ENEURO.0063-20.2020

Fodor, J. A. (1981). Representations: Philosophical Essays on the Foundations of Cognitive Science . Brighton, Sussex: The Harvester Press.

Gazzaniga, M. S., Ivry, R. B., and Mangun, G. R. (2014). Cognitive Neuroscience: The Biology of Mind (4th). New York, NY: W. W. Norton & Company Ltd.

Ghuman, A. S., Brunet, N. M., Li, Y., Konecky, R. O., Pyles, J. A., Walls, S. A., et al. (2014). Dynamic encoding of face information in the human fusiform gyrus. Nat. Commun. 5, 1–10. doi: 10.1038/ncomms6672

Greenbaum, S., and Quirk, R. (1970). Elicitation Experiments in English: Linguistic Studies in Use and Attitude . Coral Gables, FL: University of Miami Press.

Griffiths, P. E. (1997). What Emotions Really Are . Chicago, IL: University of Chicago Press.

Griffiths, P., Machery, E., and Linquist, S. (2009). The vernacular concept of innateness. Mind Lang. 24, 605–630. doi: 10.1111/j.1468-0017.2009.01376.x

Haugeland, J. (1998). Having Thought: Essays in the Metaphysics of Mind . Cambridge, MA: Harvard University Press.

Honey, C. J., and Sporns, O. (2008). Dynamical consequences of lesions in cortical networks. Hum. Brain Mapp. 29, 802–809. doi: 10.1002/hbm.20579

Hull, D. L. (1988). Science as a Process: An Evolutionary Account of the Social and Conceptual Development of Science . Chicago, IL: University of Chicago Press.

Hurford, J. R. (2007). The Origins of Meaning: Language in the Light of Evolution . New York, NY: Oxford University Press.

Hutto, D. D., and Myin, E. (2013). Radicalizing Enactivism: Basic Minds Without Content . Cambridge, MA: The MIT Press.

Izhikevich, E. (2007). Dynamical Systems in Neuroscience: The Geometry of Excitability and Bursting . Cambridge, MA: MIT Press.

Jia, Z., Ikeda, R., Ling, J., and Gu, J. G. (2013). GTP-dependent run-up of Piezo2-type mechanically activated currents in rat dorsal root ganglion neurons. Mol. Brain 6, 1–8. doi: 10.1186/1756-6606-6-57

Kiefer, M., and Pulvermüller, F. (2012). Conceptual representations in mind and brain: theoretical developments, current evidence and future directions. Cortex 48, 805–825. doi: 10.1016/j.cortex.2011.04.006

Klein, S. B. (2015). What memory is. Wiley Interdiscip. Rev. Cogn. Sci. 6, 1–38. doi: 10.1002/wcs.1333

Kriegeskorte, N., and Diedrichsen, J. (2019). Peeling the onion of brain representations. Annu. Rev. Neurosci. 42, 407–432. doi: 10.1146/annurev-neuro-080317-061906

Laird, J. E. (2012). The Soar Cognitive Architecture . Cambridge, MA: The MIT press.

Machery, E. (2009). Doing Without Concepts . New York, NY: Oxford University Press.

Machery, E. (2017). Philosophy Within Its Proper Bounds . Oxford: Oxford University Press.

Machery, E., Griffiths, P., Linquist, S., and Stotz, K. (2019). “Scientists’ concepts of innateness: evolution or attraction?” in Advances in Experimental Philosophy of Science . eds. D. A. Wilkenfeld and R. Samuels (New York, NY: Bloomsbury), 172–201.

Mahon, B. Z., and Caramazza, A. (2008). A critical look at the embodied cognition hypothesis and a new proposal for grounding conceptual content. J. Physiol. Paris 102, 59–70. doi: 10.1016/j.jphysparis.2008.03.004

Millikan, R. G. (1989). An ambiguity in the notion “function.”. Biol. Philos. 4, 172–176. doi: 10.1007/BF00127747

Piantadosi, S. T., Tenenbaum, J. B., and Goodman, N. D. (2016). The logical primitives of thought: empirical foundations for compositional cognitive models. Psychol. Rev. 123, 392–424. doi: 10.1037/a0039980

Piazza, M., and Eger, E. (2016). Neural foundations and functional specificity of number representations. Neuropsychologia 83, 257–273. doi: 10.1016/j.neuropsychologia.2015.09.025

Planer, R. J., and Godfrey-Smith, P. (2021). Communication and representation understood as sender-receiver coordination. Mind Lang. 36, 750–770. doi: 10.1111/mila.12293

Poldrack, R. A. (2021). The physics of representation. Synthese 199, 1307–1325. doi: 10.1007/s11229-020-02793-y

Ramsey, W. M. (2007). Representation Reconsidered . New York, NY: Cambridge University Press.

Reid, R. C., and Usrey, W. M. (2013). Vision. In L. R. Squire, D. Berg, F. E. Bloom, S. LacDu, A. Ghosh, and N. C. Spitzer (Eds.), Fundamentals of Neuroscience 4th, pp. 577–598. Waltham, MA: Academic Press.

Rheinberger, H.-J. (2000). “Fragments from the perspective of molecular biology” in Decoding the Genetic Program . eds. E. F. Keller, P. Beurton, R. Falk, and H.-J. Rheinberger (Cambridge, MA: Cambridge University Press), 219–239.

Richardson, M. J., Shockley, K., Fajen, B. R., Riley, M. R., and Turvey, M. T. (2008). “Ecological psychology: six principles for an embodied-embedded approach to behavior” in Handbook of Cognitive Science: An Embodied Approach . eds. R. Calvo and T. Gomila (Amsterdam: Elsevier Science), 161–187.

Roskies, A. L. (2021). Representational similarity analysis in neuroimaging: proxy vehicles and provisional representations. Synthese 199, 5917–5935. doi: 10.1007/s11229-021-03052-4

Schnotz, W., Hauck, G., and Schwartz, N. H. (2021). Multiple mental representations in picture processing. Psychol. Res. 86, 903–918. doi: 10.1007/s00426-021-01541-2

Shannon, C. E., and Weaver, W. (1949/1964) The Mathematical Theory Of Communication . Urbana, IL: The University of Illinois Press.

Shea, N. (2018). Representation in Cognitive Science . Oxford, UK: Oxford University Press.

Shenoy, K. V., Sahani, M., and Churchland, M. M. (2013). Cortical control of arm movements: a dynamical systems perspective. Annu. Rev. Neurosci. 36, 337–359. doi: 10.1146/annurev-neuro-062111-150509

Soh, Z., Sakamoto, K., Suzuki, M., Iino, Y., and Tsuji, T. (2018). A computational model of internal representations of chemical gradients in environments for chemotaxis of Caenorhabditis elegans . Sci. Rep. 8:17190. doi: 10.1038/s41598-018-35157-1

Sussillo, D., and Barak, O. (2013). Opening the black box: low-dimensional dynamics in high-dimensional recurrent neural networks. Neural Comput. 25, 626–649. doi: 10.1162/NECO_a_00409

Tarr, M. J., and Gauthier, I. (2000). FFA: a flexible fusiform area for subordinate-level visual processing automatized by expertise. Nat. Neurosci. 3, 764–769. doi: 10.1038/77666

Trappenberg, T. P. (2010). Fundamentals of Computational Neuroscience (2nd). New York, NY: Oxford University Press.

Vilarroya, O. (2017). Neural representation. A survey-based analysis of the notion. Front. Psychol. Theor. Phil. Psychol. 8:1458. doi: 10.3389/fpsyg.2017.01458

Voelker, A. R., Blouw, P., Choo, X., Dumont, N. S. Y., Stewart, T. C., and Eliasmith, C. (2021). Simulating and predicting dynamical systems with spatial semantic pointers. Neural Comput. 33, 2033–2067. doi: 10.1162/neco_a_01410

Von Eckardt, B. (1995). What is Cognitive Science? Cambridge, MA: The MIT Press.

Yablo, S. (2014). Aboutness . Princeton, NJ: Princeton University Press.

Zhang, M., Kalies, W. D., Kelso, J. S., and Tognoli, E. (2020). Topological portraits of multiscale coordination dynamics. J. Neurosci. Methods 339:108672. doi: 10.1016/j.jneumeth.2020.108672

Keywords: representation, conceptual reform, information, scientific concepts, cognition

Citation: Favela LH and Machery E (2023) Investigating the concept of representation in the neural and psychological sciences. Front. Psychol . 14:1165622. doi: 10.3389/fpsyg.2023.1165622

Received: 23 February 2023; Accepted: 19 April 2023; Published: 07 June 2023.

Reviewed by:

Copyright © 2023 Favela and Machery. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Luis H. Favela, [email protected]

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

• More from M-W
• To save this word, you'll need to log in. Log In

Definition of represent

 (Entry 1 of 2)

transitive verb

intransitive verb

Definition of re-present  (Entry 2 of 2)

• characterize

Examples of represent in a Sentence

These examples are programmatically compiled from various online sources to illustrate current usage of the word 'represent.' Any opinions expressed in the examples do not represent those of Merriam-Webster or its editors. Send us feedback about these examples.

Word History

Middle English, from Anglo-French representer , from Latin repraesentare , from re- + praesentare to present

14th century, in the meaning defined at transitive sense 1

1564, in the meaning defined above

Dictionary Entries Near represent

reprehensory

Cite this Entry

“Represent.” Merriam-Webster.com Dictionary , Merriam-Webster, https://www.merriam-webster.com/dictionary/represent. Accessed 11 Apr. 2024.

Kids Definition

Kids definition of represent, legal definition, legal definition of represent, more from merriam-webster on represent.

Nglish: Translation of represent for Spanish Speakers

Britannica English: Translation of represent for Arabic Speakers

Subscribe to America's largest dictionary and get thousands more definitions and advanced search—ad free!

Can you solve 4 words at once?

Word of the day.

See Definitions and Examples »

Get Word of the Day daily email!

Popular in Grammar & Usage

Your vs. you're: how to use them correctly, every letter is silent, sometimes: a-z list of examples, more commonly mispronounced words, how to use em dashes (—), en dashes (–) , and hyphens (-), absent letters that are heard anyway, popular in wordplay, the words of the week - apr. 5, 12 bird names that sound like compliments, 10 scrabble words without any vowels, 12 more bird names that sound like insults (and sometimes are), 8 uncommon words related to love, games & quizzes.

• Previous Chapter
• Next Chapter
• Download PDF

The concept of representation central in contemporary interpretations of democracy is in many ways dependent also from the juridical, artistic and religious languages, and the meanings it assumes in this field. This polysemic character has animated the history of political thought, where the concept of representation has been viewed in different and loosely related ways. An important turning point for the contemporary development of the scientific (and political) debate has been the formation of a consensus around the meaning of representation within the context of the neo-Schumpeterian view of democracy, in which the adjective representative referred to the influence of citizens’ opinion on policy-making. The seminal work of Hanna Pitkin shifted the focus on the substantive character of political representation conceived as acting in the interests of the represented. Both approaches were built around the concept of responsiveness, and coexisted as standard references for several decades. Around the end of the twentieth century the concept of representation and the related practices were object of a renewed attention both in response to the progress of the academic debate and as a consequence of the changing political reality.

Table of Contents

• [66.249.64.20|185.147.128.134]
• 185.147.128.134

Character limit 500 /500

• Table of Contents
• Random Entry
• Chronological
• Editorial Information
• About the SEP
• Editorial Board
• How to Cite the SEP
• Special Characters
• Advanced Tools
• Support the SEP
• PDFs for SEP Friends
• Make a Donation
• SEPIA for Libraries
• Entry Contents

Bibliography

Academic tools.

• Friends PDF Preview
• Author and Citation Info
• Back to Top

Cognitive Science

Cognitive science is the interdisciplinary study of mind and intelligence, embracing philosophy, psychology, artificial intelligence, neuroscience, linguistics, and anthropology. Its intellectual origins are in the mid-1950s when researchers in several fields began to develop theories of mind based on complex representations and computational procedures. Its organizational origins are in the mid-1970s when the Cognitive Science Society was formed and the journal Cognitive Science began. Since then, more than one hundred universities in North America, Europe, Asia, and Australia have established cognitive science programs, and many others have instituted courses in cognitive science.

3. Representation and Computation

4.1 formal logic, 4.3 concepts, 4.4 analogies, 4.6 connectionism, 4.7 theoretical neuroscience, 4.8 bayesian, 4.9 deep learning, 4.10 predictive processing and active inference, 5.1 philosophical applications, 5.2 critique of cognitive science, 5.3 philosophy of cognitive science, other internet resources, related entries.

Attempts to understand the mind and its operation go back at least to the Ancient Greeks, when philosophers such as Plato and Aristotle tried to explain the nature of human knowledge. The study of mind remained the province of philosophy until the nineteenth century, when experimental psychology developed. Wilhelm Wundt and his students initiated laboratory methods for studying mental operations more systematically. Within a few decades, however, experimental psychology became dominated by behaviorism , a view that virtually denied the existence of mind. According to behaviorists such as J. B. Watson, psychology should restrict itself to examining the relation between observable stimuli and observable behavioral responses. Talk of consciousness and mental representations was banished from respectable scientific discussion. Especially in North America, behaviorism dominated the psychological scene through the 1950s.

Around 1956, the intellectual landscape began to change dramatically. George Miller summarized numerous studies which showed that the capacity of human thinking is limited, with short-term memory, for example, limited to around seven items. He proposed that memory limitations can be overcome by recoding information into chunks, mental representations that require mental procedures for encoding and decoding the information. At this time, primitive computers had been around for only a few years, but pioneers such as John McCarthy, Marvin Minsky, Allen Newell, and Herbert Simon were founding the field of artificial intelligence . In addition, Noam Chomsky rejected behaviorist assumptions about language as a learned habit and proposed instead to explain language comprehension in terms of mental grammars consisting of rules. The six thinkers mentioned in this paragraph can be viewed as the founders of cognitive science.

Cognitive science has unifying theoretical ideas, but we have to appreciate the diversity of outlooks and methods that researchers in different fields bring to the study of mind and intelligence. Although cognitive psychologists today often engage in theorizing and computational modeling, their primary method is experimentation with human participants. People, often undergraduates satisfying course requirements, are brought into the laboratory so that different kinds of thinking can be studied under controlled conditions. For example, psychologists have experimentally examined the kinds of mistakes people make in deductive reasoning, the ways that people form and apply concepts, the speed of people thinking with mental images, and the performance of people solving problems using analogies. Our conclusions about how the mind works must be based on more than “common sense” and introspection, since these can give a misleading picture of mental operations, many of which are not consciously accessible. Increasingly, psychologists draw their experimental participants from Amazon’s Mechanical Turk and from culturally diverse sources. Psychological experiments that carefully approach mental operations from diverse directions are therefore crucial for cognitive science to be scientific. Experimentation is also a methodology employed by experimental philosophy.

Although theory without experiment is empty, experiment without theory is blind. To address the crucial questions about the nature of mind, the psychological experiments need to be interpretable within a theoretical framework that postulates mental representations and procedures. One of the best ways of developing theoretical frameworks is by forming and testing computational models intended to be analogous to mental operations. To complement psychological experiments on deductive reasoning, concept formation, mental imagery, and analogical problem solving, researchers have developed computational models that simulate aspects of human performance. Designing, building, and experimenting with computational models is the central method of artificial intelligence (AI), the branch of computer science concerned with intelligent systems. Ideally in cognitive science, computational models and psychological experimentation go hand in hand, but much important work in AI has examined the power of different approaches to knowledge representation in relative isolation from experimental psychology.

While some linguists do psychological experiments or develop computational models, most currently use different methods. For linguists in the Chomskian tradition, the main theoretical task is to identify grammatical principles that provide the basic structure of human languages. Identification takes place by noticing subtle differences between grammatical and ungrammatical utterances. In English, for example, the sentences “She hit the ball” and “What do you like?” are grammatical, but “She the hit ball” and “What does you like?” are not. A grammar of English will explain why the former are acceptable but not the latter. An alternative approach, cognitive linguistics, puts less emphasis on syntax and more on semantics and concepts.

Like cognitive psychologists, neuroscientists often perform controlled experiments, but their observations are very different, since neuroscientists are concerned directly with the nature of the brain. With nonhuman subjects, researchers can insert electrodes and record the firing of individual neurons. With humans for whom this technique would be too invasive, it is now common to use magnetic and positron scanning devices to observe what is happening in different parts of the brain while people are doing various mental tasks. For example, brain scans have identified the regions of the brain involved in mental imagery and word interpretation. Additional evidence about brain functioning is gathered by observing the performance of people whose brains have been damaged in identifiable ways. A stroke, for example, in a part of the brain dedicated to language can produce deficits such as the inability to utter sentences. Like cognitive psychology, neuroscience is often theoretical as well as experimental, and theory development is frequently aided by developing computational models of the behavior of groups of neurons.

Cognitive anthropology expands the examination of human thinking to consider how thought works in different cultural settings. The study of mind should obviously not be restricted to how English speakers think but should consider possible differences in modes of thinking across cultures. Cognitive science is becoming increasingly aware of the need to view the operations of mind in particular physical and social environments. For cultural anthropologists, the main method is ethnography, which requires living and interacting with members of a culture to a sufficient extent that their social and cognitive systems become apparent. Cognitive anthropologists have investigated, for example, the similarities and differences across cultures in words for colors.

Traditionally, philosophers do not perform systematic empirical observations or construct computational models, although there has been a rise in work in experimental philosophy. But philosophy remains important to cognitive science because it deals with fundamental issues that underlie the experimental and computational approach to mind. Abstract questions such as the nature of representation and computation need not be addressed in the everyday practice of psychology or artificial intelligence, but they inevitably arise when researchers think deeply about what they are doing. Philosophy also deals with general questions such as the relation of mind and body and with methodological questions such as the nature of explanations found in cognitive science. In addition, philosophy concerns itself with normative questions about how people should think as well as with descriptive ones about how they do. Besides the theoretical goal of understanding human thinking, cognitive science can have the practical goal of improving it, which requires normative reflection on what we want thinking to be. Philosophy of mind does not have a distinct method, but should share with the best theoretical work in other fields a concern with empirical results.

In its weakest form, cognitive science is just the sum of the fields mentioned: psychology, artificial intelligence, linguistics, neuroscience, anthropology, and philosophy. Interdisciplinary work becomes much more interesting when there is theoretical and experimental convergence on conclusions about the nature of mind. For example, psychology and artificial intelligence can be combined through computational models of how people behave in experiments. The best way to grasp the complexity of human thinking is to use multiple methods, especially psychological and neurological experiments and computational models. Theoretically, the most fertile approach has been to understand the mind in terms of representation and computation.

The central hypothesis of cognitive science is that thinking can best be understood in terms of representational structures in the mind and computational procedures that operate on those structures. While there is much disagreement about the nature of the representations and computations that constitute thinking, the central hypothesis is general enough to encompass the current range of thinking in cognitive science, including connectionist theories which model thinking using artificial neural networks.

Most work in cognitive science assumes that the mind has mental representations analogous to computer data structures, and computational procedures similar to computational algorithms. Cognitive theorists have proposed that the mind contains such mental representations as logical propositions, rules, concepts, images, and analogies, and that it uses mental procedures such as deduction, search, matching, rotating, and retrieval. The dominant mind-computer analogy in cognitive science has taken on a novel twist from the use of another analog, the brain.

Connectionists have proposed novel ideas about representation and computation that use neurons and their connections as inspirations for data structures, and neuron firing and spreading activation as inspirations for algorithms. Cognitive science then works with a complex 3-way analogy among the mind, the brain, and computers. Mind, brain, and computation can each be used to suggest new ideas about the others. There is no single computational model of mind, since different kinds of computers and programming approaches suggest different ways in which the mind might work. The computers that most of us work with today are serial processors, performing one instruction at a time, but the brain and some recently developed computers are parallel processors, capable of doing many operations at once.

A major trend in current cognitive science is the integration of neuroscience with many areas of psychology, including cognitive, social, developmental, and clinical. This integration is partly experimental, resulting from an explosion of new instruments for studying the brain, such as functional magnetic resonance imaging, transcranial magnetic stimulation, and optogenetics. The integration is also theoretical, because of advances in understanding how large populations of neurons can perform tasks usually explained with cognitive theories of rules and concepts.

4. Theoretical Approaches

Here is a schematic summary of current theories about the nature of the representations and computations that explain how the mind works.

Formal logic provides some powerful tools for looking at the nature of representation and computation. Propositional and predicate calculus serve to express many complex kinds of knowledge, and many inferences can be understood in terms of logical deduction with inferences rules such as modus ponens. The explanation schema for the logical approach is:

Explanation target: Why do people make the inferences they do? Explanatory pattern: People have mental representations similar to sentences in predicate logic. People have deductive and inductive procedures that operate on those sentences. The deductive and inductive procedures, applied to the sentences, produce the inferences.

It is not certain, however, that logic provides the core ideas about representation and computation needed for cognitive science, since more efficient and psychologically natural methods of computation may be needed to explain human thinking. (See the entry on logic and artificial intelligence .)

Much of human knowledge is naturally described in terms of rules of the form IF … THEN …, and many kinds of thinking such as planning can be modeled by rule-based systems. The explanation schema used is:

Explanation target: Why do people have a particular kind of intelligent behavior? Explanatory pattern: People have mental rules. People have procedures for using these rules to search a space of possible solutions, and procedures for generating new rules. Procedures for using and forming rules produce the behavior.

Computational models based on rules have provided detailed simulations of a wide range of psychological experiments, from cryptarithmetic problem solving to skill acquisition to language use. Rule-based systems have also been of practical importance in suggesting how to improve learning and how to develop intelligent machine systems.

Concepts, which partly correspond to the words in spoken and written language, are an important kind of mental representation. There are computational and psychological reasons for abandoning the classical view that concepts have strict definitions. Instead, concepts can be viewed as sets of typical features. Concept application is then a matter of getting an approximate match between concepts and the world. Schemas and scripts are more complex than concepts that correspond to words, but they are similar in that they consist of bundles of features that can be matched and applied to new situations. The explanatory schema used in concept-based systems is:

Explanatory target: Why do people have a particular kind of intelligent behavior? Explanation pattern: People have a set of concepts, organized via kind and part hierarchies and other associations. People have a set of procedures for concept application, including spreading activation, matching, and inheritance. The procedures applied to the concepts produce the behavior. Concepts can be translated into rules, but they bundle information differently than sets of rules, making possible different computational procedures.

Analogies play an important role in human thinking, in areas as diverse as problem solving, decision making, explanation, and linguistic communication. Computational models simulate how people retrieve and map source analogs in order to apply them to target situations. The explanation schema for analogies is:

Explanation target: Why do people have a particular kind of intelligent behavior? Explanatory pattern: People have verbal and visual representations of situations that can be used as cases or analogs. People have processes of retrieval, mapping, and adaptation that operate on those analogs. The analogical processes, applied to the representations of analogs, produce the behavior.

The constraints of similarity, structure, and purpose overcome the difficult problem of how previous experiences can be found and used to help with new problems. Not all thinking is analogical, and using inappropriate analogies can hinder thinking, but analogies can be effective in applications such as education and design.

Visual and other kinds of images play an important role in human thinking. Pictorial representations capture visual and spatial information in a much more usable form than lengthy verbal descriptions. Computational procedures well suited to visual representations include inspecting, finding, zooming, rotating, and transforming. Such operations can be very useful for generating plans and explanations in domains to which pictorial representations apply. The explanatory schema for visual representation is:

Explanation target: Why do people have a particular kind of intelligent behavior? Explanatory pattern: People have visual images of situations. People have processes such as scanning and rotation that operate on those images. The processes for constructing and manipulating images produce the intelligent behavior.

Imagery can aid learning, and some metaphorical aspects of language may have their roots in imagery. Psychological experiments suggest that visual procedures such as scanning and rotating employ imagery, and neurophysiological results confirm a close physical link between reasoning with mental imagery and perception. Imagery is not just visual, but can also operate with other sensory experiences such as hearing, touch, smell, taste, pain, balance, nausea, fullness, and emotion.

Connectionist networks consisting of simple nodes and links are very useful for understanding psychological processes that involve parallel constraint satisfaction. Such processes include aspects of vision, decision making, explanation selection, and meaning making in language comprehension. Connectionist models can simulate learning by methods that include Hebbian learning and backpropagation. The explanatory schema for the connectionist approach is:

Explanation target: Why do people have a particular kind of intelligent behavior? Explanatory pattern: People have representations that involve simple processing units linked to each other by excitatory and inhibitory connections. People have processes that spread activation between the units via their connections, as well as processes for modifying the connections. Applying spreading activation and learning to the units produces the behavior.

Simulations of various psychological experiments have shown the psychological relevance of the connectionist models, which are, however, only very rough approximations to actual neural networks. (For more information, see the entry on connectionism .)

Theoretical neuroscience is the attempt to develop mathematical and computational theories and models of the structures and processes of the brains of humans and other animals. It differs from connectionism in trying to be more biologically accurate by modeling the behavior of large numbers of realistic neurons organized into functionally significant brain areas. Computational models of the brain have become biologically richer, both with respect to employing more realistic neurons such as ones that spike and have chemical pathways, and with respect to simulating the interactions among different areas of the brain such as the hippocampus and the cortex. These models are not strictly an alternative to computational accounts in terms of logic, rules, concepts, analogies, images, and connections, but should mesh with them and show how mental functioning can be performed at the neural level. The explanatory schema for theoretical neuroscience is:

Explanation target: How does the brain carry out functions such as cognitive tasks? Explanatory pattern: The brain has neurons organized by synaptic connections into populations and brain areas. The neural populations have spiking patterns that are transformed via sensory inputs and the spiking patterns of other neural populations. Interactions of neural populations carry out functions including cognitive tasks.

From the perspective of theoretical neuroscience, mental representations are patterns of neural activity, and inference is transformation of such patterns. (See the entries on neuroscience and the neuroscience of consciousness .)

Bayesian models are prominent in cognitive science, with applications to such psychological phenomena as learning, vision, motor control, language, and social cognition. They have also had effective applications in robotics. The Bayesian approach assumes that cognition is approximately optimal in accord with probability theory, especially Bayes’ theorem, which says that the probability of a hypothesis given evidence is equal to the result of multiplying the prior probability of the hypothesis by the conditional probability of the evidence given the hypothesis, all divided by the probability of the evidence. The explanatory schema for Bayesian cognition is:

Explanation target: How does the mind carry out functions such as inference? Explanatory pattern: The mind has representations for statistical correlations and conditional probabilities. The mind has the capacity for probabilistic computations such as applications of Bayes’ theorem. Applying probabilistic computations to statistical representations accomplishes mental tasks such as inference.

Although Bayesian methods have had impressive applications to a wide range of phenomena, their psychological plausibility is debatable because of assumptions about optimality and computations based on probability theory.

Artificial intelligence has been a central part of cognitive since the 1950s, and the most dramatic recent advances in AI have come from the approach of deep learning, which has produced major breakthroughs in fields that include game playing, object recognition, and translation. Deep learning builds on ideas from connectionism and theoretical neuroscience, but uses neural networks with more layers and improved algorithms, benefitting from faster computers and large data bases of examples. Another important innovation is combining learning from examples with reinforcement learning, resulting by 2016 in the world’s leading Go player, AlphaGo. Ideas from deep learning are spreading back into neuroscience and also beginning to influence research in cognitive psychology. The explanatory schema for deep learning is:

Explanation target: How does the brain carry out functions such as cognitive tasks? Explanatory pattern: The brain has large numbers of neurons organized into 6–20 layers. The brain has powerful mechanisms for learning from examples and for learning actions that are reinforced by their successes. Applying learning mechanisms to layered neural networks makes them capable of human and sometimes even super-human performance.

Although deep learning has produced dramatic improvements in some AI systems, it is not clear how it can be applied to aspects of human thought that include causal reasoning, imagery, emotion, and analogy. For further discussion, see Section 11 (on deep learning) of the entry on connectionism .

Predictive processing is an approach to theoretical neuroscience that views the brain as constantly generating and updating models of the environment in order to predict the results of perceptions and actions. Active inference is a version of predictive processing that hypothesizes that the brain uses Bayesian calculations to minimize “free energy” consisting of discrepancies between expectations and actual observations. Organisms survive when brains reduce prediction errors by changing their models of the environment or by changing the environment through action.

The explanatory schema for active inference is:

Explanation target: How does the brain function to support perception and action? Explanatory pattern: The brain is a prediction engine that uses probabilistic models to anticipate perceptions and the results of actions. To reduce prediction error, the brain uses Bayesian updating to change its models and uses actions to change its environment, e.g. by moving. Effective inference, perception, and action result from these reductions in prediction errors.

Active inference is open to numerous challenges. Is brain functioning really Bayesian updating rather than connectionist constraint satisfaction or deep reinforcement learning? Can predictive processing subsume other brain functions that include pattern recognition, explanation, emotional evaluation, memory, and communication? Does active inference explain high-level cognitive operations such as causal reasoning, language, and creativity?

5. Philosophical Relevance

Some philosophy, in particular naturalistic philosophy of mind, is part of cognitive science. But the interdisciplinary field of cognitive science is relevant to philosophy in several ways. First, the psychological, computational, and other results of cognitive science investigations have important potential applications to traditional philosophical problems in epistemology, metaphysics, and ethics. Second, cognitive science can serve as an object of philosophical critique, particularly concerning the central assumption that thinking is representational and computational. Third and more constructively, cognitive science can be taken as an object of investigation in the philosophy of science, generating reflections on the methodology and presuppositions of the enterprise.

Much philosophical research today is naturalistic, treating philosophical investigations as continuous with empirical work in fields such as psychology. From a naturalistic perspective, philosophy of mind is closely allied with theoretical and experimental work in cognitive science. Metaphysical conclusions about the nature of mind are to be reached, not by a priori speculation, but by informed reflection on scientific developments in fields such as psychology, neuroscience, and computer science. Similarly, epistemology is not a stand-alone conceptual exercise, but depends on and benefits from scientific findings concerning mental structures and learning procedures. Ethics can benefit by using greater understanding of the psychology of moral thinking to bear on ethical questions such as the nature of deliberations concerning right and wrong. Here are some philosophical problems to which ongoing developments in cognitive science are highly relevant. Links are provided to other relevant articles in this Encyclopedia.

• Innateness . To what extent is knowledge innate or acquired by experience? Is human behavior shaped primarily by nature or nurture?
• Language of thought . Does the human brain operate with a language-like code or with a more general connectionist architecture? What is the relation between symbolic cognitive models using rules and concepts and sub-symbolic models using neural networks?
• Mental imagery . Do human minds think with visual and other kinds of imagery, or only with language-like representations?
• Folk psychology . Does a person’s everyday understanding of other people consist of having a theory of mind, or of merely being able to simulate them?
• Meaning . How do mental representations acquire meaning or mental content ? To what extent does the meaning of a representation depend on its relation to other representations, its relation to the world, and its relation to a community of thinkers?
• Mind-brain identity . Are mental states brain states? Or can they be multiply realized by other material states? What is the relation between psychology and neuroscience? Is materialism true?
• Free will . Is human action free or merely caused by brain events?
• Moral psychology . How do minds/brains make ethical judgments?
• The meaning of life . How can minds construed naturalistically as brains find value and meaning?
• Emotions . What are emotions, and what role do they play in thinking?
• Consciousness . Can conscious experience be scientifically explained, for example by the neuroscience of consciousness ?
• Mental disorder . What are mental disorders, and how are psychological and neural processes relevant to their explanation and treatment?
• Perception and reality . How do minds/brains form and evaluate representations of the external world?
• Perception and cognition . How does perception differ from other kinds of cognition with respect to representational format and justification?
• Realism . Is cognitive science consistent with views that minds grasp the real world? Could minds be computer simulations? Is virtual reality a kind of reality?
• Information . How does cognitive science illuminate the operations of information and misinformation in minds and societies?
• Social science . How do explanations of the operations of minds interact with explanations of the operations of groups and societies?

Additional philosophical problems arise from examining the presuppositions of current approaches to cognitive science.

The claim that human minds work by representation and computation is an empirical conjecture and might be wrong. Although the computational-representational approach to cognitive science has been successful in explaining many aspects of human problem solving, learning, and language use, some philosophical critics have claimed that this approach is fundamentally mistaken. Critics of cognitive science have offered such challenges as:

• The emotion challenge: Cognitive science neglects the important role of emotions in human thinking.
• The consciousness challenge: Cognitive science ignores the importance of consciousness in human thinking.
• The world challenge: Cognitive science disregards the significant role of physical environments in human thinking, which is embedded in and extended into the world.
• The body challenge: Cognitive science neglects the contribution of embodiment to human thought and action.
• The dynamical systems challenge: The mind is a dynamical system, not a computational system.
• The social challenge: Human thought is inherently social in ways that cognitive science ignores.
• The mathematics challenge: Mathematical results show that human thinking cannot be computational in the standard sense, so the brain must operate differently, perhaps as a quantum computer.
• The interdisciplinarity challenge: Cognitive science has failed to go beyond multidisciplinary interactions by developing a core theory that unifies work in its many disciplines.

The first five challenges are increasingly addressed by advances that explain emotions, consciousness, action, and embodiment in terms of neural mechanisms. The social challenge is being met by the development of computational models of interacting agents. The mathematics challenge is based on misunderstanding of Gödel’s theorem and on exaggeration of the relevance of quantum theory to neural processes. Response to the interdisciplinary challenge must recognize that cognitive science still has many contending theoretical approaches, without the unification that theories of evolution and genetics provide for biology. Nevertheless, interactions among psychology, neuroscience, linguistics, philosophy, anthropology, and computer modeling have contributed to theoretical and empirical progress concerning many aspects of cognition. For example, computational philosophy uses programmed models to address questions in epistemology, ethics, and other areas of philosophy.

Cognitive science raises many interesting methodological questions that are worthy of investigation by philosophers of science. What is the nature of representation? What role do computational models play in the development of cognitive theories? What is the relation among apparently competing accounts of mind involving symbolic processing, neural networks, and dynamical systems? What is the relation among the various fields of cognitive science such as psychology, linguistics, and neuroscience? Are psychological phenomena subject to reductionist explanations via neuroscience? Are levels of explanation best characterized in terms of ontological levels (molecular, neural, psychological, social) or methodological ones (computational, algorithmic, physical)?

The increasing prominence of neural explanations in cognitive, social, developmental, and clinical psychology raises important philosophical questions about explanation and reduction . Anti-reductionism, according to which psychological explanations are completely independent of neurological ones, is becoming increasingly implausible, but it remains controversial to what extent psychology can be reduced to neuroscience and molecular biology. Crucial to answering questions about the nature of reduction are answers to questions about the nature of explanation. Explanations in psychology, neuroscience, and biology in general are plausibly viewed as descriptions of mechanisms , which are combinations of connected parts that interact to produce regular changes. In psychological explanations, the parts are mental representations that interact by computational procedures to produce new representations. In neuroscientific explanations, the parts are neural populations that interact by electrochemical processes to produce new neural activity that leads to actions. If progress in theoretical neuroscience continues, it should become possible to tie psychological to neurological explanations by showing how mental representations such as concepts are constituted by activities in neural populations, and how computational procedures such as spreading activation among concepts are carried out by neural processes.

The increasing integration of cognitive psychology with neuroscience provides evidence for the mind-brain identity theory according to which mental processes are neural, representational, and computational. Other philosophers dispute such identification on the grounds that minds are embodied in biological systems and extended into the world. However, moderate claims about embodiment are consistent with the identity theory because brain representations operate in several modalities (e.g. visual and motor) that enable minds to deal with the world. Another materialist alternative to mind-brain identity comes from recognizing that explanations of mind employ molecular and social mechanisms as well as neural and representational ones.

• Anderson, J. R., 2007. How Can the Mind Occur in the Physical Universe? , Oxford: Oxford University Press.
• –––, 2010. Cognitive Psychology and its Implications , 7th edition, New York: Worth.
• Bechtel, W., 2008. Mental Mechanisms: Philosophical Perspectives on Cognitive Neurosciences , New York: Routledge.
• Bechtel, W., & Graham, G. (eds.), 1998. A Companion to Cognitive Science , Malden, MA: Blackwell.
• Bechtel, W., Mandik, P., Mundale, J., & Stufflebeam, R. S. (eds.), 2001. Philosophy and the Neurosciences: A Reader , Malden, MA: Blackwell.
• Bermúdez, J. L., 2022. Cognitive Science: An Introduction to the Science of the Mind , 4th edition, Cambridge: Cambridge University Press.
• Blouw, P., Solodkin, E., Thagard, P., & Eliasmith, C., 2016. “Concepts as Semantic Pointers: A Framework and Computational Model,” Cognitive Science , 40: 1128–1162.
• Chalmers, D. J., 2022. Reality+: Virtual Worlds and the Problems of Philosophy , New York: Norton.
• Boden, M. A., 2006. Mind as Machine: A History of Cognitive Science , Oxford: Clarendon.
• Chemero, A., 2009. Radical Embodied Cognitive Science , Cambridge, MA: MIT Press.
• Churchland, P. M., 2007. Neurophilosophy at Work , Cambridge: Cambridge University Press.
• Churchland, P. S., 2002. Brain-wise: Studies in Neurophilosophy , Cambridge, MA: MIT Press.
• Clark, A., 2001. Mindware: An Introduction to the Philosophy of Cognitive science , New York: Oxford University Press.
• –––, 2008. Supersizing the Mind: Embodiment, Action, and Cognitive Extension , New York: Oxford University Press.
• –––, 2015. Surfing Uncertainty: Prediction, Action, and the Embodied mind , Oxford: Oxford University Press.
• Craver, C. F., 2007. Explaining the Brain , Oxford: Oxford University Press.
• Dayan, P., & Abbott, L. F., 2001. Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems , Cambridge, MA: MIT Press.
• Dehaene, S., 2014. Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts , New York: Viking.
• Doya, K., Ishii, S., Pouget, A., & Rao, A. (eds.), 2007. Bayesian Brain , Cambridge, MA: MIT Press.
• Dreyfus, H. L., 1992. What Computers Still Can’t Do , 3rd edition, Cambridge, MA: MIT Press.
• Eliasmith, C., 2013. How to Build a Brain: A Neural Architecture for Biological Cognition , Oxford: Oxford University Press.
• Eliasmith, C., & Anderson, C. H., 2003. Neural Engineering: Computation, Representation and Dynamics in Neurobiological Systems , Cambridge, MA: MIT Press.
• Forbus, K. D., Ferguson, R. W., Lovett, A., & Gentner, D., 2017. “Extending SME to Handle Larger-Scale Cognitive modeling,”. Cognitive Science , 41(5): 1152–1201.
• Friedenberg, J. D., & Silverman, G., 2021. Cognitive Science: An Introduction to the Study of Mind , 4th edition, Thousand Oaks, CA: Sage.
• Gibbs, R. W., 2005, Embodiment and Cognitive Science , Cambridge: Cambridge University Press.
• Goldman, A., 1993. Philosophical Applications of Cognitive Science , Boulder: Westview Press.
• Goodfellow, I., Bengio, Y., & Courville, A., 2016. Deep Learning , Cambridge, MA: MIT Press.
• Griffiths, T. L., Kemp, C., & Tenenbaum, J. B., 2008. “Bayesian Models of Cognition,” in R. Sun (ed.), The Cambridge Handbook of Computational Psychology , Cambridge: Cambridge University Press, pp. 59–100.
• Hoffman, D., 2019. The Case against Reality: Why Evolution Hid the Truth From our Eyes , New York: WW Norton & Company.
• Hofstadter, D., & Sander, E., 2013. Surfaces and Essences: Analogy as the Fuel and Fire of Thinking , New York: Basic Books.
• Holyoak, K. J., & Morrison, R. G. (eds.), 2012. The Oxford Handbook of Thinking and Reasoning , New York: Oxford University Press.
• Knobe, J., & Nichols, S. (eds.), 2008. Experimental Philosophy , Oxford: Oxford University Press.
• Kallens, P. C., Dale, R., & Christiansen, M. H., 2022. “Quantifying Interdisciplinarity in Cognitive Science and Beyond.” Topics in Cognitive Science , 14(3): 634–645.
• Kosslyn, S. M., Thompson, W. L., & Ganis, G., 2016. The Case for Mental Imagery , New York: Oxford University Press.
• Laird, J. E., LeBiere, C., & Rosenbloom, P. S., 2017. “A Standard Model of the Mind: Toward a Common Computational Framework across Artificial Intelligence, Cognitive Science, Neuroscience, and Robotics,” AI Magazine , 38(4): 13–26.
• Lecun, Y., Bengio, Y., & Hinton, Y., 2015. “Deep Learning,” Nature , 521(7553): 436–444.
• Margolis, E. & Laurence, S. (eds.), 2015. The Conceptual Mind: New Directions in the Study of Concepts , Cambridge, MA: MIT Press.
• McCauley, R. N., 2007. “Reduction: Models of Cross-scientific Relations and their Implications for the Psychology-neuroscience Interface,” in P. Thagard (ed.), Philosophy of Psychology and Cognitive Science , Amsterdam: Elsevier, pp. 105–158.
• Milkowski, M., 2013. Explaining the Computational Mind , Cambridge, MA: MIT Press.
• Murphy, D., 2006. Psychiatry in the Scientific Image , Cambridge, MA: MIT Press.
• Nadel, L. (ed.), 2003. Encyclopedia of Cognitive Science , London:Nature Publishing Group.
• Nisbett, R., 2003. The Geography of Thought: How Asians and Westerners Think Differently … and Why , New York: Free Press.
• Núñez, R., Allen, M., Gao, R., Miller Rigoli, C., Relaford-Doyle, J., & Semenuks, A., 2019. “What Happened to Cognitive Science?” Nature Human Behavior , 3(8): 782–791.
• O’Callaghan, C. A Multisensory Philosophy of Perception . Oxford: Oxford University of Press.
• O’Reilly, R. C., Munakata, Y., Frank, M. J., Hazy, T. E., & Contributors, 2012. Computational Cognitive Neuroscience , Wiki Book, [ O’Reilly et al. 2012 available online ].
• Parr, T., Pezzulo, G., & Friston, K. J., 2022. Active Inference: The Free Energy Principle in Mind, Brain, and Behaviour , Cambridge, MA: MIT Press.
• Pearl, J., & Mackenzie, D., 2018. The Book of Why: The New Science of Cause and Effect , New York: Basic Books.
• Pessoa, L., 2013. The Cognitive-Emotional Brain: From Interactions to Integration , Cambridge, MA: MIT Press.
• Piantadosi, S. T., Tenenbaum, J. B., & Goodman, N. D., 2015. “The Logical Primitives of Thought: Empirical Foundations for Compositional Cognitive Models,”. Psychological Review , 123(4), 392–424.
• Polk, T. A., & Seifert, C. M. (eds.), 2002. Cognitive Modeling , Cambridge, MA: MIT Press.
• Quilty-Dunn, J., 2020. “Perceptual pluralism,” Noûs , 54(4): 807–838.
• Rogers, T. T., & McClelland, J. L., 2004. Semantic Cognition: A Parallel Distributed Processing Approach , Cambridge, MA: MIT Press.
• Russell, S., & Norvig, P., 2020. Artificial Intelligence: A Modern Approach , 4th edition, New York: Pearson.
• Seth, A., 2021. Being You: A New Science of Consciousness , New York: Dutton
• Smith, E. E.., & Kosslyn, S. M., 2007. Cognitive Psychology: Mind and Brain , Upper Saddle River, NJ: Pearson Prentice Hall.
• Sun, R. (ed.), 2008. The Cambridge Handbook of Computational Psychology , Cambridge: Cambridge University Press.
• ––– (ed.), 2012. Grounding Social Sciences in Cognitive Sciences , Cambridge, MA: MIT Press.
• Thagard, P., 2005. Mind: Introduction to Cognitive Science , second edition, Cambridge, MA: MIT Press.
• –––, 2009. “Why cognitive science needs philosophy and vice versa, ” Topics in Cognitive Science , 1: 237–254.
• –––, 2010. The Brain and the Meaning of Life , Princeton: Princeton University Press.
• –––, 2012. The Cognitive Science of Science: Explanation, Discovery, and Conceptual Change , Cambridge, MA: MIT Press.
• –––, 2019a. Brain-Mind: From Neurons to Consciousness and Creativity , Oxford: Oxford University Press.
• –––, 2019b. Natural Philosophy: From Social Brains to Knowledge, Reality, Morality, and Beauty , New York: Oxford University Press.
• –––, 2021. Bots and Beasts: What Makes Machines, Animals, and People Smart? Cambridge, MA: MIT Press.
• Thompson, E., 2007. Mind in Life: Biology, Phenomenology, and the Science of Mind , Cambridge, MA: Harvard University Press.
• Varela, F. J., Thompson, E., & Rosch, E., 2016. The Embodied Mind: Cognitive Science and Human Experience , revised edition, Cambridge, MA: MIT Press.
• Wilson, R. A., & Keil, F. C. (eds.), 1999. The MIT Encyclopedia of the Cognitive Sciences , Cambridge, MA: MIT Press.

How to cite this entry . Preview the PDF version of this entry at the Friends of the SEP Society . Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO). Enhanced bibliography for this entry at PhilPapers , with links to its database.

• Artificial Intelligence on the Web
• Biographies of Major Contributors to Cognitive Science
• Cognitive Science Dictionary, University of Alberta
• Cognitive Science Society
• Dictionary of Philosophy of Mind
• Glossary of Cognitive Science
• Mind and Brain News from Science Daily

artificial intelligence | artificial intelligence: logic-based | behaviorism | cognition: embodied | computational philosophy | concepts | connectionism | consciousness | consciousness: neuroscience of | emotion | experimental philosophy | folk psychology: as a theory | folk psychology: as mental simulation | free will | innate/acquired distinction | innateness: and contemporary theories of cognition | intentionality | language of thought hypothesis | learning, perceptual | meaning, theories of | memory | mental content: causal theories of | mental disorder | mental imagery | mental representation | mind/brain identity theory | mind: computational theory of | mind: modularity of | moral psychology: empirical approaches | neuroscience, philosophy of | perception: the contents of

Copyright © 2023 by Paul Thagard < pthagard @ uwaterloo . ca >

• Accessibility

Support SEP

Mirror sites.

View this site from another server:

• Info about mirror sites

The Stanford Encyclopedia of Philosophy is copyright © 2023 by The Metaphysics Research Lab , Department of Philosophy, Stanford University

Library of Congress Catalog Data: ISSN 1095-5054

• Search Menu
• Browse content in Arts and Humanities
• Browse content in Archaeology
• Anglo-Saxon and Medieval Archaeology
• Archaeological Methodology and Techniques
• Archaeology by Region
• Archaeology of Religion
• Archaeology of Trade and Exchange
• Biblical Archaeology
• Contemporary and Public Archaeology
• Environmental Archaeology
• Historical Archaeology
• History and Theory of Archaeology
• Industrial Archaeology
• Landscape Archaeology
• Mortuary Archaeology
• Prehistoric Archaeology
• Underwater Archaeology
• Urban Archaeology
• Zooarchaeology
• Browse content in Architecture
• Architectural Structure and Design
• History of Architecture
• Residential and Domestic Buildings
• Theory of Architecture
• Browse content in Art
• Art Subjects and Themes
• History of Art
• Industrial and Commercial Art
• Theory of Art
• Biographical Studies
• Byzantine Studies
• Browse content in Classical Studies
• Classical History
• Classical Philosophy
• Classical Mythology
• Classical Literature
• Classical Reception
• Classical Art and Architecture
• Classical Oratory and Rhetoric
• Greek and Roman Epigraphy
• Greek and Roman Law
• Greek and Roman Papyrology
• Greek and Roman Archaeology
• Late Antiquity
• Religion in the Ancient World
• Digital Humanities
• Browse content in History
• Colonialism and Imperialism
• Diplomatic History
• Environmental History
• Genealogy, Heraldry, Names, and Honours
• Genocide and Ethnic Cleansing
• Historical Geography
• History by Period
• History of Emotions
• History of Agriculture
• History of Education
• History of Gender and Sexuality
• Industrial History
• Intellectual History
• International History
• Labour History
• Legal and Constitutional History
• Local and Family History
• Maritime History
• Military History
• National Liberation and Post-Colonialism
• Oral History
• Political History
• Public History
• Regional and National History
• Revolutions and Rebellions
• Slavery and Abolition of Slavery
• Social and Cultural History
• Theory, Methods, and Historiography
• Urban History
• World History
• Browse content in Language Teaching and Learning
• Language Learning (Specific Skills)
• Language Teaching Theory and Methods
• Browse content in Linguistics
• Applied Linguistics
• Cognitive Linguistics
• Computational Linguistics
• Forensic Linguistics
• Grammar, Syntax and Morphology
• Historical and Diachronic Linguistics
• History of English
• Language Acquisition
• Language Evolution
• Language Reference
• Language Variation
• Language Families
• Lexicography
• Linguistic Anthropology
• Linguistic Theories
• Linguistic Typology
• Phonetics and Phonology
• Psycholinguistics
• Sociolinguistics
• Translation and Interpretation
• Writing Systems
• Browse content in Literature
• Bibliography
• Children's Literature Studies
• Literary Studies (Asian)
• Literary Studies (European)
• Literary Studies (Eco-criticism)
• Literary Studies (Romanticism)
• Literary Studies (American)
• Literary Studies (Modernism)
• Literary Studies - World
• Literary Studies (1500 to 1800)
• Literary Studies (19th Century)
• Literary Studies (20th Century onwards)
• Literary Studies (African American Literature)
• Literary Studies (British and Irish)
• Literary Studies (Early and Medieval)
• Literary Studies (Fiction, Novelists, and Prose Writers)
• Literary Studies (Gender Studies)
• Literary Studies (Graphic Novels)
• Literary Studies (History of the Book)
• Literary Studies (Plays and Playwrights)
• Literary Studies (Poetry and Poets)
• Literary Studies (Postcolonial Literature)
• Literary Studies (Queer Studies)
• Literary Studies (Science Fiction)
• Literary Studies (Travel Literature)
• Literary Studies (War Literature)
• Literary Studies (Women's Writing)
• Literary Theory and Cultural Studies
• Mythology and Folklore
• Shakespeare Studies and Criticism
• Browse content in Media Studies
• Browse content in Music
• Applied Music
• Dance and Music
• Ethics in Music
• Ethnomusicology
• Gender and Sexuality in Music
• Medicine and Music
• Music Cultures
• Music and Religion
• Music and Media
• Music and Culture
• Music Education and Pedagogy
• Music Theory and Analysis
• Musical Scores, Lyrics, and Libretti
• Musical Structures, Styles, and Techniques
• Musicology and Music History
• Performance Practice and Studies
• Race and Ethnicity in Music
• Sound Studies
• Browse content in Performing Arts
• Browse content in Philosophy
• Aesthetics and Philosophy of Art
• Epistemology
• Feminist Philosophy
• History of Western Philosophy
• Metaphysics
• Moral Philosophy
• Non-Western Philosophy
• Philosophy of Science
• Philosophy of Language
• Philosophy of Mind
• Philosophy of Perception
• Philosophy of Action
• Philosophy of Law
• Philosophy of Religion
• Philosophy of Mathematics and Logic
• Practical Ethics
• Social and Political Philosophy
• Browse content in Religion
• Biblical Studies
• Christianity
• East Asian Religions
• History of Religion
• Judaism and Jewish Studies
• Qumran Studies
• Religion and Education
• Religion and Health
• Religion and Politics
• Religion and Science
• Religion and Law
• Religion and Art, Literature, and Music
• Religious Studies
• Browse content in Society and Culture
• Cookery, Food, and Drink
• Cultural Studies
• Customs and Traditions
• Ethical Issues and Debates
• Hobbies, Games, Arts and Crafts
• Lifestyle, Home, and Garden
• Natural world, Country Life, and Pets
• Popular Beliefs and Controversial Knowledge
• Sports and Outdoor Recreation
• Technology and Society
• Travel and Holiday
• Visual Culture
• Browse content in Law
• Arbitration
• Browse content in Company and Commercial Law
• Commercial Law
• Company Law
• Browse content in Comparative Law
• Systems of Law
• Competition Law
• Browse content in Constitutional and Administrative Law
• Government Powers
• Judicial Review
• Local Government Law
• Military and Defence Law
• Parliamentary and Legislative Practice
• Construction Law
• Contract Law
• Browse content in Criminal Law
• Criminal Procedure
• Criminal Evidence Law
• Sentencing and Punishment
• Employment and Labour Law
• Environment and Energy Law
• Browse content in Financial Law
• Banking Law
• Insolvency Law
• History of Law
• Human Rights and Immigration
• Intellectual Property Law
• Browse content in International Law
• Private International Law and Conflict of Laws
• Public International Law
• IT and Communications Law
• Jurisprudence and Philosophy of Law
• Law and Politics
• Law and Society
• Browse content in Legal System and Practice
• Courts and Procedure
• Legal Skills and Practice
• Primary Sources of Law
• Regulation of Legal Profession
• Medical and Healthcare Law
• Browse content in Policing
• Criminal Investigation and Detection
• Police and Security Services
• Police Procedure and Law
• Police Regional Planning
• Browse content in Property Law
• Personal Property Law
• Study and Revision
• Terrorism and National Security Law
• Browse content in Trusts Law
• Wills and Probate or Succession
• Browse content in Medicine and Health
• Browse content in Allied Health Professions
• Arts Therapies
• Clinical Science
• Dietetics and Nutrition
• Occupational Therapy
• Operating Department Practice
• Physiotherapy
• Radiography
• Speech and Language Therapy
• Browse content in Anaesthetics
• General Anaesthesia
• Neuroanaesthesia
• Browse content in Clinical Medicine
• Acute Medicine
• Cardiovascular Medicine
• Clinical Genetics
• Clinical Pharmacology and Therapeutics
• Dermatology
• Endocrinology and Diabetes
• Gastroenterology
• Genito-urinary Medicine
• Geriatric Medicine
• Infectious Diseases
• Medical Toxicology
• Medical Oncology
• Pain Medicine
• Palliative Medicine
• Rehabilitation Medicine
• Respiratory Medicine and Pulmonology
• Rheumatology
• Sleep Medicine
• Sports and Exercise Medicine
• Clinical Neuroscience
• Community Medical Services
• Critical Care
• Emergency Medicine
• Forensic Medicine
• Haematology
• History of Medicine
• Browse content in Medical Dentistry
• Oral and Maxillofacial Surgery
• Paediatric Dentistry
• Restorative Dentistry and Orthodontics
• Surgical Dentistry
• Browse content in Medical Skills
• Clinical Skills
• Communication Skills
• Nursing Skills
• Surgical Skills
• Medical Ethics
• Medical Statistics and Methodology
• Browse content in Neurology
• Clinical Neurophysiology
• Neuropathology
• Nursing Studies
• Browse content in Obstetrics and Gynaecology
• Gynaecology
• Occupational Medicine
• Ophthalmology
• Otolaryngology (ENT)
• Browse content in Paediatrics
• Neonatology
• Browse content in Pathology
• Chemical Pathology
• Clinical Cytogenetics and Molecular Genetics
• Histopathology
• Medical Microbiology and Virology
• Patient Education and Information
• Browse content in Pharmacology
• Psychopharmacology
• Browse content in Popular Health
• Caring for Others
• Complementary and Alternative Medicine
• Self-help and Personal Development
• Browse content in Preclinical Medicine
• Cell Biology
• Molecular Biology and Genetics
• Reproduction, Growth and Development
• Primary Care
• Professional Development in Medicine
• Browse content in Psychiatry
• Addiction Medicine
• Child and Adolescent Psychiatry
• Forensic Psychiatry
• Learning Disabilities
• Old Age Psychiatry
• Psychotherapy
• Browse content in Public Health and Epidemiology
• Epidemiology
• Public Health
• Browse content in Radiology
• Clinical Radiology
• Interventional Radiology
• Nuclear Medicine
• Radiation Oncology
• Reproductive Medicine
• Browse content in Surgery
• Cardiothoracic Surgery
• Gastro-intestinal and Colorectal Surgery
• General Surgery
• Neurosurgery
• Paediatric Surgery
• Peri-operative Care
• Plastic and Reconstructive Surgery
• Surgical Oncology
• Transplant Surgery
• Trauma and Orthopaedic Surgery
• Vascular Surgery
• Browse content in Science and Mathematics
• Browse content in Biological Sciences
• Aquatic Biology
• Biochemistry
• Bioinformatics and Computational Biology
• Developmental Biology
• Ecology and Conservation
• Evolutionary Biology
• Genetics and Genomics
• Microbiology
• Molecular and Cell Biology
• Natural History
• Plant Sciences and Forestry
• Research Methods in Life Sciences
• Structural Biology
• Systems Biology
• Zoology and Animal Sciences
• Browse content in Chemistry
• Analytical Chemistry
• Computational Chemistry
• Crystallography
• Environmental Chemistry
• Industrial Chemistry
• Inorganic Chemistry
• Materials Chemistry
• Medicinal Chemistry
• Mineralogy and Gems
• Organic Chemistry
• Physical Chemistry
• Polymer Chemistry
• Study and Communication Skills in Chemistry
• Theoretical Chemistry
• Browse content in Computer Science
• Artificial Intelligence
• Computer Architecture and Logic Design
• Game Studies
• Human-Computer Interaction
• Mathematical Theory of Computation
• Programming Languages
• Software Engineering
• Systems Analysis and Design
• Virtual Reality
• Browse content in Computing
• Business Applications
• Computer Security
• Computer Games
• Computer Networking and Communications
• Digital Lifestyle
• Graphical and Digital Media Applications
• Operating Systems
• Browse content in Earth Sciences and Geography
• Atmospheric Sciences
• Environmental Geography
• Geology and the Lithosphere
• Maps and Map-making
• Meteorology and Climatology
• Oceanography and Hydrology
• Palaeontology
• Physical Geography and Topography
• Regional Geography
• Soil Science
• Urban Geography
• Browse content in Engineering and Technology
• Agriculture and Farming
• Biological Engineering
• Civil Engineering, Surveying, and Building
• Electronics and Communications Engineering
• Energy Technology
• Engineering (General)
• Environmental Science, Engineering, and Technology
• History of Engineering and Technology
• Mechanical Engineering and Materials
• Technology of Industrial Chemistry
• Transport Technology and Trades
• Browse content in Environmental Science
• Applied Ecology (Environmental Science)
• Conservation of the Environment (Environmental Science)
• Environmental Sustainability
• Environmentalist Thought and Ideology (Environmental Science)
• Management of Land and Natural Resources (Environmental Science)
• Natural Disasters (Environmental Science)
• Nuclear Issues (Environmental Science)
• Pollution and Threats to the Environment (Environmental Science)
• Social Impact of Environmental Issues (Environmental Science)
• History of Science and Technology
• Browse content in Materials Science
• Ceramics and Glasses
• Composite Materials
• Metals, Alloying, and Corrosion
• Nanotechnology
• Browse content in Mathematics
• Applied Mathematics
• Biomathematics and Statistics
• History of Mathematics
• Mathematical Education
• Mathematical Finance
• Mathematical Analysis
• Numerical and Computational Mathematics
• Probability and Statistics
• Pure Mathematics
• Browse content in Neuroscience
• Cognition and Behavioural Neuroscience
• Development of the Nervous System
• Disorders of the Nervous System
• History of Neuroscience
• Invertebrate Neurobiology
• Molecular and Cellular Systems
• Neuroendocrinology and Autonomic Nervous System
• Neuroscientific Techniques
• Sensory and Motor Systems
• Browse content in Physics
• Astronomy and Astrophysics
• Atomic, Molecular, and Optical Physics
• Biological and Medical Physics
• Classical Mechanics
• Computational Physics
• Condensed Matter Physics
• Electromagnetism, Optics, and Acoustics
• History of Physics
• Mathematical and Statistical Physics
• Measurement Science
• Nuclear Physics
• Particles and Fields
• Plasma Physics
• Quantum Physics
• Relativity and Gravitation
• Semiconductor and Mesoscopic Physics
• Browse content in Psychology
• Affective Sciences
• Clinical Psychology
• Cognitive Psychology
• Cognitive Neuroscience
• Criminal and Forensic Psychology
• Developmental Psychology
• Educational Psychology
• Evolutionary Psychology
• Health Psychology
• History and Systems in Psychology
• Music Psychology
• Neuropsychology
• Organizational Psychology
• Psychological Assessment and Testing
• Psychology of Human-Technology Interaction
• Psychology Professional Development and Training
• Research Methods in Psychology
• Social Psychology
• Browse content in Social Sciences
• Browse content in Anthropology
• Anthropology of Religion
• Human Evolution
• Medical Anthropology
• Physical Anthropology
• Regional Anthropology
• Social and Cultural Anthropology
• Theory and Practice of Anthropology
• Browse content in Business and Management
• Business Strategy
• Business Ethics
• Business History
• Business and Government
• Business and Technology
• Business and the Environment
• Comparative Management
• Corporate Governance
• Corporate Social Responsibility
• Entrepreneurship
• Health Management
• Human Resource Management
• Industrial and Employment Relations
• Industry Studies
• Information and Communication Technologies
• International Business
• Knowledge Management
• Management and Management Techniques
• Operations Management
• Organizational Theory and Behaviour
• Pensions and Pension Management
• Public and Nonprofit Management
• Strategic Management
• Supply Chain Management
• Browse content in Criminology and Criminal Justice
• Criminal Justice
• Criminology
• Forms of Crime
• International and Comparative Criminology
• Youth Violence and Juvenile Justice
• Development Studies
• Browse content in Economics
• Agricultural, Environmental, and Natural Resource Economics
• Asian Economics
• Behavioural Finance
• Behavioural Economics and Neuroeconomics
• Econometrics and Mathematical Economics
• Economic Systems
• Economic History
• Economic Methodology
• Economic Development and Growth
• Financial Markets
• Financial Institutions and Services
• General Economics and Teaching
• Health, Education, and Welfare
• History of Economic Thought
• International Economics
• Labour and Demographic Economics
• Law and Economics
• Macroeconomics and Monetary Economics
• Microeconomics
• Public Economics
• Urban, Rural, and Regional Economics
• Welfare Economics
• Browse content in Education
• Adult Education and Continuous Learning
• Care and Counselling of Students
• Early Childhood and Elementary Education
• Educational Equipment and Technology
• Educational Strategies and Policy
• Higher and Further Education
• Organization and Management of Education
• Philosophy and Theory of Education
• Schools Studies
• Secondary Education
• Teaching of a Specific Subject
• Teaching of Specific Groups and Special Educational Needs
• Teaching Skills and Techniques
• Browse content in Environment
• Applied Ecology (Social Science)
• Climate Change
• Conservation of the Environment (Social Science)
• Environmentalist Thought and Ideology (Social Science)
• Natural Disasters (Environment)
• Social Impact of Environmental Issues (Social Science)
• Browse content in Human Geography
• Cultural Geography
• Economic Geography
• Political Geography
• Browse content in Interdisciplinary Studies
• Communication Studies
• Museums, Libraries, and Information Sciences
• Browse content in Politics
• African Politics
• Asian Politics
• Chinese Politics
• Comparative Politics
• Conflict Politics
• Elections and Electoral Studies
• Environmental Politics
• European Union
• Foreign Policy
• Gender and Politics
• Human Rights and Politics
• Indian Politics
• International Relations
• International Organization (Politics)
• International Political Economy
• Irish Politics
• Latin American Politics
• Middle Eastern Politics
• Political Methodology
• Political Communication
• Political Philosophy
• Political Sociology
• Political Behaviour
• Political Economy
• Political Institutions
• Political Theory
• Politics and Law
• Public Administration
• Public Policy
• Quantitative Political Methodology
• Regional Political Studies
• Russian Politics
• Security Studies
• State and Local Government
• UK Politics
• US Politics
• Browse content in Regional and Area Studies
• African Studies
• Asian Studies
• East Asian Studies
• Japanese Studies
• Latin American Studies
• Middle Eastern Studies
• Native American Studies
• Scottish Studies
• Browse content in Research and Information
• Research Methods
• Browse content in Social Work
• Addictions and Substance Misuse
• Adoption and Fostering
• Care of the Elderly
• Child and Adolescent Social Work
• Couple and Family Social Work
• Developmental and Physical Disabilities Social Work
• Direct Practice and Clinical Social Work
• Emergency Services
• Human Behaviour and the Social Environment
• International and Global Issues in Social Work
• Mental and Behavioural Health
• Social Justice and Human Rights
• Social Policy and Advocacy
• Social Work and Crime and Justice
• Social Work Macro Practice
• Social Work Practice Settings
• Social Work Research and Evidence-based Practice
• Welfare and Benefit Systems
• Browse content in Sociology
• Childhood Studies
• Community Development
• Comparative and Historical Sociology
• Economic Sociology
• Gender and Sexuality
• Gerontology and Ageing
• Health, Illness, and Medicine
• Marriage and the Family
• Migration Studies
• Occupations, Professions, and Work
• Organizations
• Population and Demography
• Race and Ethnicity
• Social Theory
• Social Movements and Social Change
• Social Research and Statistics
• Social Stratification, Inequality, and Mobility
• Sociology of Religion
• Sociology of Education
• Sport and Leisure
• Urban and Rural Studies
• Browse content in Warfare and Defence
• Defence Strategy, Planning, and Research
• Land Forces and Warfare
• Military Administration
• Military Life and Institutions
• Naval Forces and Warfare
• Other Warfare and Defence Issues
• Peace Studies and Conflict Resolution
• Weapons and Equipment

• < Previous chapter
• Next chapter >

Hanna Pitkin, The Concept of Representation

School of Government and Public Policy, University of Arizona

• Published: 10 December 2015
• Cite Icon Cite
• Permissions Icon Permissions

This chapter reviews Hanna Pitkin’s seminal book, The Concept of Representation , her most important and lasting contribution to political philosophy. Using Ludwig Wittgenstein’s ordinary-language theory, Pitkin explores the semantic landscape and the etymology of the concept of representation. In particular, she expounds on the meaning of representation by proposing four different views of representation: formalistic representation, descriptive representation, symbolic representation, and substantive representation. This chapter discusses Pitkin’s arguments, with particular emphasis on her assertion that representation is a paradox and that genuine representation respects the autonomy of the represented and the representative. It also considers the limitations of The Concept of Representation , such as its failure to adequately examine the relationship between democracy and representation.

Hanna Pitkin’s brilliance is in her meticulous and methodical attention to the underlying structure and meaning of political concepts. She excavates past meanings, traces their etymological development, and surveys the terrain of contemporary usages, not to solve political problems, but to uncover and understand the tensions and paradoxes that constitute political life. Pitkin systemically presents the language through which humans describe, understand, and evaluate political behavior in order to show how people can have different and contradictory aims. Her work also highlights the costs for pursuing any one aim in a singular fashion. In this way, Pitkin’s work provides a conceptual roadmap for navigating the vast and puzzling nature of political ideas.

By far, Hanna Pitkin’s most important and lasting contribution to political philosophy is The Concept of Representation . Using Ludwig Wittgenstein’s ordinary-language theory, this seminal book sketches the semantic landscape and the etymology of the concept of representation. Pitkin details how a map can “represent” a local topography, an actor can “represent” Hamlet, an attorney can “represent” a litigant, and a senator can “represent” her constituents. Despite these seemingly disparate usages, Pitkin identifies a common meaning: Representation is “the making present of something which is nevertheless not literally present” (143). Notice the paradox: representation requires both being present and being not present. Pitkin traces this paradox throughout various usages of representation. For instance, this paradox is most apparent in the activity of representation: representatives should be responsive to their constituents’ preferences (constituents’ preferences present) and act independently in accordance with their constituents’ best interests (constituents’ preferences not present). Her paradoxical approach to representation focuses on whether the autonomy of both the represented and the representative is being preserved. Pitkin maintains that genuine representation requires both. For this reason, Pitkin does not try to reconcile these paradoxical features of representation; rather, she illuminates these contradictions in order to reduce misunderstandings. For Pitkin, political life entails many paradoxical and contradictory commitments. An important task of political theory is to help humans identify and clarify these paradoxes.

The Concept of Representation usefully catalogues the vocabulary used to describe, explicate, and assess electoral behavior. Furthermore, Pitkin’s insistence that the meaning of representation depends on its usage enables her analysis to apply to political behavior beyond elections: Pitkin’s catalogue can pertain to non-state actors such as international, transnational, and non-governmental organizations as well as to courts, interest groups, lobbyists, and social movements. As the meaning of representation has become inextricably tied to modern notions of democracy, the significance of Pitkin’s analysis has grown. Her work lays out the underlying schematic foundation of democratic representative institutions and practices. Pitkin’s contribution is enduring because it can accommodate changing political practices and thereby incorporate the evolving meanings of representation.

Pitkin’s assertions, that representation is a paradox and that genuine representation respects the autonomy of the represented and the representative, provide grounds for criticizing different forms of governance such as illiberal democracies ( Zakaria 1997 ) and hybrid regimes ( Diamond 2002 ). After all, in her discussion of how kings can be representatives and how the Soviet Union holds elections, Pitkin recognizes that representative institutions can increase the legitimacy of tyrants and serve undemocratic purposes. Her paradoxical understanding of representation, though, offers reasons for criticizing even the deliberative turns of authoritarian regimes, for example, China, if deliberations on the local level do not translate into sufficient capacities to act on the central level ( He and Warren 2011 ).

Furthermore, Pitkin’s etymological analysis of fascist forms of representation holds implicit warnings to those who want to eliminate the paradoxical nature of representation: Representation can become mere authority, when representative institutions suppress the autonomy of the represented. Even the responsiveness of representatives to citizens’ preferences can be used as a form of control. In this way, Pitkin’s schematic understanding of representation provides reasons for criticizing and even rejecting certain representative practices.

That said, Pitkin recognizes that there are multiple ways to negotiate the contradictory standards for representation. For her, the activity of representation is necessarily pluralist. Because genuine representation requires representatives to act independently and in a manner responsive to the represented, representatives can balance these competing standards in different acceptable ways. For this reason, Pitkin holds that appropriate standards for evaluating representation will depend importantly on the nature of interests, welfare, and wishes of the represented, the relative capacities of representative and constituents, and the nature of the issues with which the representative must deal (210). In this way, Pitkin recommends shifting the theoretical attention away from general questions such as “how should an individual represent?” to questions focused on how a particular individual should act in a particular institutional and political context.

So to understand democratic representation using Pitkin’s approach, one needs to ask the following questions: What are the appropriate standards for democratic representation? 1 What are the interests of democratic citizens that should guide and constrain their representatives’ behavior? Do representatives and their constituents have the necessary capacities for democratic representation? What is the nature of the issues facing democratic representatives and their constituents? While her theoretical framework recommends asking certain questions, Pitkin does not provide answers to them.

For instance, Pitkin never adequately specifies how we are to identify constituents’ objective interests. At times, she implies that constituents should have “some say” in what these are. However, she recognizes that trustees will sometimes need to advance constituents’ interests by going against their preferences. For Pitkin, representatives who do so must only provide justifications for their actions. It is unclear, though, how or whether “genuine” representative processes should guarantee or require representatives to give such explanations. Perhaps it follows that transparency should be understood as an “objective interest” of democratic citizens, but Pitkin does not elaborate.

Instead, Pitkin concludes with the somewhat unsatisfying formulation that representatives “must not be found persistently at odds with the wishes of the represented without good reason in terms of their interest, without a good explanation of why their wishes are not in accord with their interest”(210). By qualifying her position with the word “persistently,” Pitkin suggests it is only the frequency of decisions contrary to constituents’ preferences without adequate explanation that prevents genuine representation. In this way, Pitkin’s general theoretical approach to evaluating representatives thwarts making particular evaluations. Pitkin does not provide absolute principles that should be used as benchmarks for evaluating a representative’s performance; rather, she offers guidelines about which kinds of questions we should pursue. And for Pitkin, answers to political problems depend in large part on the kind of question being asked.

For example, Pitkin (1965 , 1966 ) saw the “problem of political obligation” and of whether citizens should obey the law as arising from four different clusters of questions. The first cluster, which she calls “the limits of obligation,” considers “when is a person obligated to obey and when not?” This cluster seeks to establish the conditions that necessitate an obligation. The second cluster focuses on “the locus of sovereignty”—that is, the question of “whom are you obligated to obey?” This recognizes that the state is not a monolithic structure; rather, the public sphere can produce various conflicting obligations. The third cluster examines the difference between legitimate authority and mere coercion. In other words, some questions about political obligation are really about the proper use of force. The final cluster investigates the justifications of obligation: “Why are you ever obligated to obey even a legitimate authority?” Pitkin argues that these various clusters of questions generate distinct and sometimes incompatible answers. An answer to the question “whom you should obey” will not provide a satisfactory solution to the question “why should you obey a legitimate authority?” Similarly, a question “why should you obey a legitimate authority” will not inform you how you should differentiate a legitimate authority from mere coercion. Given the different issues implicit in the problem of political obligation, answers to the question “should a person obey the law” will depends in a large part which cluster of questions is being stressed (1965, 990).

In addition to revealing the need to clarify political questions, Pitkin catalogues the vocabulary with which we investigate and assess representative processes and practices within democracies. Two things strike me as remarkable about using Pitkin’s analysis of representation as the schematic foundation for understanding democratic representation. First, The Concept of Representation never adequately examines the relationship between democracy and representation. The relationship is presumed, not explicated. In fact, as mentioned before, her discussion of representation includes analysis of non-democratic forms of governance. Pitkin notes that “every government claims to represent,” implying that representation is not an exclusively democratic practice. Moreover, she stresses that through much of human history, “both the concept and the practice of representation have had little to do with democracy or liberty” (2). For instance, Pitkin depicts representation as initially a burden placed on aristocrats by the king as an efficient way to collect taxes. Aristocrats were appointed as the king’s delegates. However, the institutionalization of these delegates created a political body that could eventually resist the king’s power. With this institutional transformation, representation became a matter of rights, as opposed to merely a burden.

In her later years, Pitkin admitted that The Concept of Representation took the relationship between democracy and representation “for granted as unproblematic. ” Pitkin explained that “Like most people even today, I more or less equated democracy with representation or at least with representative government” (2004, 336). Here she acknowledges how contemporary understandings of a political concept can bias theoretical and conceptual analysis. Pitkin goes on to say that the conflation of democracy and representation is “profoundly misleading,” for it obscures the way “that representation increasingly supplants democracy instead of serving it” (2004, 339). Here Pitkin raises an interesting question about whether democracy can be adequately protected by a political representation that is only intermittently bolstered by citizen participation, deliberation, and enforcement of human and constitutional rights ( Castiglione 2012 ). Pitkin remains suspicious of the elitist nature of representation.

Second, her theoretical contribution to assessing representative processes in democracies is remarkable given her claim that the concept of representation is, and will always be , a paradox. Satisfying the criteria of representation demanded by one meaning of representation requires violating the criteria required by other meanings. Paradoxes might be descriptively accurate, capturing complexities and contradictions generated by various usages of the term; however, their explanatory power is limited to the degree that they ascribe multiple interpretations to the same phenomena. How does one argue with, let alone refute, a paradox? Pitkin sketches the paradoxical terrain of representation with little advice about how to navigate the normative complexities and ambiguities of representation. In the end, Pitkin requires her readers to use their own judgment for assessing the compromises and bargaining that constitute and accompany representative processes.

The Landscape of Representation

How did Pitkin sketch and catalogue the various contradictory meanings of representation? She famously captures the enigmatic nature of representation in a metaphor. She envisions the concept of representation as “a rather complicated, convoluted, three-dimensional structure in the middle of a dark enclosure.” Political theorists provide only “flash-bulb photographs of the structure taken from different angles” (10). In other words, political theorists approach the concept of representation with necessarily limited and partial perspectives. According to Pitkin’s metaphor, political theorists are unable to “see” simultaneously the whole underlying structure of representation. What Pitkin’s metaphor does not stress, though, is how her understanding of “language as action” suggests that the three-dimensional structure can shift, grow, or shrink with common usages.

In particular, Pitkin clarifies the meaning of representation by differentiating four different views of representation. They are formalistic representation, descriptive representation, symbolic representation , and substantive representation . Each view highlights distinctive features of representation and thereby different parameters for identifying and evaluating representation.

For instance, formalistic representation examines the process of authorization and accountability. Formalistic representation considers how a representative comes to obtain his standing, status, position, or office, as well as the institutional mechanisms that encourage responsiveness to the represented. According to this view, institutional procedures, rules, and norms structure the actions of a representative. They determine how a representative can and does act. Using this view of representation, a person becomes a representative simply in virtue of having a certain job title or office. The formalistic view emphasizes the institutional procedures by which an agent acquires the authority to act.

Interestingly, Pitkin denies that formalistic representation provides any standards for assessing the actions of a representative. Formalistic representation is a kind of normative black hole. A person either has the authority to act or she does not. Of course, one can assess a representative by whether her authority was legitimately obtained or whether she exceeded her legitimate authority. But Pitkin denies that formalistic representation provides any standards for assessing how well a representative behaves. Formalistic representation is a “transaction that takes place at the outset, before the actual representing begins” (39). In this way, Pitkin downplays the ways that representatives can alter and expand their institutional authority. Consequently, her analysis does not provide any standards for assessing violations of accountability, let alone for manipulating authorization devices. Ironically, the view of representation most closely tied to accountability offers no ethical criteria for judging the performance of the representative.

Pitkin’s second view of representation, descriptive representation , focuses on the resemblance or correspondence between the representative and the represented. Of course, what should correspond can vary: Descriptive representation can focus on whether representatives’ experiences, identities, perspectives, and interests are similar to those of the represented. It is, as Anne Phillips famously described, focused on “a politics of presence” that examines who is literally participating in political processes. So, descriptive representation can refer to having representatives from urban and rural areas as well as having representatives who belong to historically marginalized groups.

Pitkin helpfully enumerates various metaphors used to describe the nature of this resemblance. Legislatures can be a picture, a mirror, a portrait, a map, a blueprint, a miniature, or even a sample of the citizenry body as a whole. Each of these different metaphors provides different criteria for judging the correspondence between the represented and the representative. For instance, one might judge a map by whether the resemblance allows one to navigate more effectively while one could judge a portrait by its ability to capture the model aesthetically. Despite such variations, Pitkin maintains that the proper way of evaluating descriptive representation is the accuracy of the correspondence or resemblance between the representative and represented. When we evaluate descriptive representatives, we judge them by whether they adequately look like or share relevant opinions, experiences, and interests of those they represent.

In her discussion of descriptive representation, Pitkin provocatively claims that increasing descriptive representation is likely to decrease accountability to that group. Pitkin based her claim on A. Phillip Griffith’s and Richard Wollheim’s provocative argument (1960) that although lunatics might be the best descriptive representatives of other lunatics, it is not desirable to send lunatics to the legislature. For Pitkin, assessing representation in accordance with the accuracy of the resemblance between representatives and represented distracts attention away from the importance of what representatives do. So judging representatives by whether they “look like” the represented weakens judging them by how well representatives advance their interests and opinions. These controversial claims made by Pitkin sparked an entire empirical and theoretical literature about whether and how descriptive representation of marginalized groups matter. According to Anne Phillips, Pitkin’s The Concept of Representation was both an “inspiration and foil” to the normative arguments used to justify increasing the representation of women in legislatures and decision-making assemblies. Phillips explained, “It was inspiration because it forced us to think more carefully about the meanings we were attributing to ‘underrepresentation.’ It was also, however, foil because it seemed so discouragingly critical of descriptive representation. It figured, therefore, as the position that had to be argued down” (2012, 513).

The third view of representation, symbolic representation , captures the ways that a representative “stands for” the represented. To help clarify this concept, consider the example of a king. The presence of a king can occasion certain patriotic emotional responses toward the constitutional monarchy. The king embodies the political significance of the kingdom as a whole. Such emotional responses reflect an implicit consent toward a king’s authority. For this reason, a king is more likely “to stand for” the country during a highly ceremonial function, such as a coronation, than during highly partisan politics. According to Pitkin, the reaction of the represented is constitutive of symbolic representation. In particular, what matters is whether a person feels represented. Such feelings will depend importantly on the attitudes and beliefs of the represented. For Pitkin, the representative is merely a passive vehicle for these emotional responses. The feelings generated by the king are like those generated by a flag.

Understood in this way, a “real” representative cannot and should not rely on propaganda or coercion to generate consent (109). Pitkin firmly differentiates accepting and following a leader from accepting the leader as a symbol of the nation. Because the latter relies on the self-understandings of citizens, it cannot easily be coerced. Such coercion would change the meaning of the representative for the represented. Pitkin acknowledges that symbolic representation often rests “on emotional, affective, irrational psychological responses rather than on rationally justifiable criteria” (100); however, she denies that representatives should actively try to shape the represented’s beliefs and identities (110). To do so would be evidence of fascist representation, not of liberal representation that respects the autonomy of the represented.

Here Pitkin’s discussion of symbolic representation appears in conflict with recent theoretical work on representation. For example, Michael Saward (2006) endorses understanding representation as a kind of claim-making. He argues that we should not understand representatives as responding to their constituents’ interests and preferences; rather, their claim-making is responsible for shaping the identities, preferences, and interests of the represented. Saward stresses how representatives “read in” interests as opposed to “read[ing] off” the interests of their constituents (2010, 310). While Saward rejects Pitkin’s framework for presuming that the represented have a “given, transparent, and relatively stable” set of interests, Pitkin could respond that dissolving the distinction between the representative and the represented is dangerous to the degree that it undermines the autonomy of the represented. It also possibly overestimates the autonomy of representatives.

The final and arguably most important view of representation, what Pitkin calls substantive representation , investigates the activities of representatives. Substantive representation refers to the behavior of acting on behalf of, in the interest of, as an agent of, or as a substitute for the represented. For Pitkin, the paradoxical requirement for being both present and not present “is precisely what appears in the mandate and independence theorists’ conflicting views on the meaning of representation” ( Pitkin 1968 , 41). The mandate–independence controversy focuses on the question, “Ought a representative to do what his constituents want, or what he thinks best?” Mandate theorists endorse a “delegate” vision of representation in which the representative is bound by the preferences of constituents. In contrast, mandate theorists endorse a “trustee” conception of representation according to which representatives should act in accordance to what they consider to be constituents’ best interests. Because interests do not necessarily align with preferences, Pitkin concludes that these standards for judging the performance of representatives’ activities are sometimes irreconcilable. Under certain circumstances, a representative will violate the standards of representation, no matter how he or she acts.

How do these various views of representation provide the theoretical vocabulary for describing and understanding democratic representative processes? If one adopts a formalistic view of representation, empirical political scientists and political theorists studying democracy focus on the institutional procedures used to authorize representatives and hold them accountable. Symbolic representation concentrates on the emotional reaction of those being represented. For example, political scientists who study symbolic representation investigate whether democratic citizens feel represented by their individual representatives, interest groups, their political parties, or formal political institutions. In contrast, descriptive representation examines whether those participating in representative processes sufficiently resemble those being represented. Here what matters is whether the representatives adequately “mirror” or “correspond” to those being represented. Substantive representation critically evaluates democratic representative processes by the criteria of whose preferences and whose interests are advanced. Together Pitkin’s various views of representation provide the conceptual framework for investigating both empirically and theoretically democratic behavior.

Following Pitkin, democratic theorists should not try to reconcile the paradoxical nature of representation since the concept of representation works properly when the autonomy of the representative is adequately balanced by the autonomy of the represented. If the representative has all the power, the system is dictatorial. If the represented have all the power, representatives are merely the mouthpieces for the mob. Pitkin wants autonomy for both. Preserving the paradoxical elements of representation is necessary for democratic representation to the extent that it safeguards both citizens and representatives. Pitkin’s analysis constrains the proper behavior of democratic representatives because they must act in ways that protect the capacity of the represented to hold them accountable. Of course, as Pitkin notes, the proper balance will depend importantly on the interests, issues, and capacities at stake in a particular democracy. Pitkin’s theoretical framework for understanding democratic representation allows for the possibility that the meaning of representation can change. But it requires preserving humans’ ability to understand and create new political practices consistent with the problems they face. Pitkin’s theoretical claim that the meaning of representation depends on its use suggests that democratic possibilities are partially created by how we understand and employ language of democratic representation.

Some contemporary political theorists resist Pitkin’s understanding of representation as a paradox. Some pose alternative “common meanings.” For instance, Andrew Rehfeld argues that the concept of representation is not paradoxical because representation can be identified “simply by reference to a relevant audience accepting a person as such” (2006, 1). Others reject Pitkin’s terminology, especially the language of interests, because it leads us to irresolvable disputes. We are highly unlikely to agree about what people’s interests are ( Celis et al. 2014 ).

Perhaps the paradoxical elements of representation could be reduced if we understood how the different views of representation fit together. If Pitkin only had given us a lexical ordering for the different views, for example, symbolic representation trumps formalistic representation, her work would provide clearer tools for assessing representative processes. Unfortunately, Pitkin never offers such an explanation. At times, she implies that the concept of representation is unified. At other times, she emphasizes the tensions among these different views. Once again, Pitkin leaves the task of putting together the various pieces of representation to us.

Limitations of The Concept of Representation

Despite its theoretical resiliency and adaptability, Pitkin’s analysis of representation might strike some as out of date. Her analysis of political representation centers on electoral relationships and frames them primarily as principal–agent problems. Such an understanding of representation is valuable to the extent that it allows constituents to settle conflicts about the proper behavior of representatives by appealing to procedural solutions. After all, elections enable constituents to both vote in and vote out their public officials. Electoral representation can simultaneously establish the legitimacy of democratic authority and create institutional incentives for governments to be responsive to citizens. Unfortunately, not all political institutions link authorization to accountability. By not recognizing how mechanisms of accountability differ significantly from those of authorization, Pitkin focuses almost exclusively on methods of authorization, specifically elections.

Moreover, political representation can often occur outside of electoral relationships ( Warren and Castiglione 2004 ). After all, no one votes for Amnesty International to speak on the behalf of the tortured. Similarly, it is misleading to think of interest groups such as American Association of Retired People as authorized by the votes of their members. As Ruth Grant and Robert Keohane (2005) claim, authorization mechanisms are often decoupled from accountability mechanisms in the international arena. By depicting representation as a principal–agent relationship, Pitkin downplays the importance of horizontal accountability ( O’Donnell 1998 )—that is, the liberal and republican components of polyarchies that allow state agencies to monitor and enforce law violations. Instead of understanding freedom as protected by the autonomy of constituents, civil and political liberties might be better protected by representative institutions with networks of agencies that enforce the rule of law. Conceiving representation as a principal–agent problem also downplays the important role of surrogate accountability-holders ( Rubenstein 2007 ) who sanction agents on the behalf of those affected by their actions. Pitkin also overlooks how representative relationships can be mediated by multi-leveled vertical relationships, for example, bureaucracies.

All of these omissions reveal a significant limitation of Pitkin’s work for understanding representation and more specifically for analyzing democratic representation: Pitkin’s analysis of accountability and its relationship to representation is woefully underdeveloped. Pitkin implicitly assumes that elections provide sufficient accountability and responsiveness to democratic citizens to guarantee their autonomy. One explanation for this assumption is that Pitkin does not consider the accountability view of representation to be “an important strand in the literature of representation” (55). Perhaps this devaluation of accountability results from the lack of any single theorist who analyzes the importance of accountability in any sustained and systematic fashion, as Thomas Hobbes did for authorization. Regardless of the reason, Pitkin perceives accountability theories as only a response and a corrective to poor authorization mechanisms. She claims that accountability theorists simply identify a representative as someone who will have to answer to another for what he does. Accountability is, therefore, a matter of responsibility and answerability. It places “new and special” obligations onto the representative.

Insightfully, Pitkin notes that we often develop our theories of democratic representative government by juxtaposing them to other (non-democratic) forms of government. Given this approach to understanding democratic representation, it is no surprise that competitive elections often emerge as “the answer” to the question, “how should representative processes be structured?” As David Plotke (1997) noted, definitions of democracy are deeply influenced by ideological commitments. During the Cold War, Joseph Schumpeter’s definition of democracy as competitive elections was useful partly because it could differentiate Western liberal democracies from Eastern European socialist countries that often identified themselves as “people’s democracies.” Pitkin’s observation about the fact that our intuitions about representative democracy can change is partially instructive because democratic representative forms of governance can share certain institutional features with non-democratic representative ones.

Recall that Pitkin maintained that formal mechanisms of authorization and accountability do not provide standards for evaluating the performance of individual representatives. They only determine whether the representative is properly authorized. Pitkin’s methodology calls for making political judgments about the appropriate standards of representation, the interests of the represented that should guide and constrain their representatives’ behavior, the necessary capacities for functioning within certain institutional setting, and the nature of the issues at stake within a particular context. The proper political language for describing and evaluating representative processes is significantly contingent upon the political norms and practices within a particular setting.

Nevertheless, Pitkin recognizes special obligations of representatives and that accountability mechanisms can justifiably constrain representatives’ behavior. For Pitkin, “genuine representation exists only where there are such controls—accountability to the represented” (57). To the degree that elections can serve to reinforce an intolerant majority’s power against minorities, or even serve merely to reinforce the power of the state, then having elections would be insufficient for both formalistic representation and democratic representation.

Pitkin does acknowledge “other forms” of accountability besides elections, yet she does not investigate how these alternative forms might influence our understanding of accountability. For instance, she never differentiates accountability mechanisms designed “to give an account of,” from accountability mechanisms designed “to hold to account” (55). In this way, she conflates transparency mechanisms with sanctioning ones. Nor does Pitkin adequately theorize the etymological distinctions between accountability that provides control and accountability that provides self-correction. Pitkin’s too-brief discussion of accountability ignores the possible paradoxical elements within it. For example, she does not explore the tensions between responsiveness and sanctioning that can arise when constituents’ interests conflict with their expressed preferences. Pitkin’s analysis of accountability is further limited by its failure to consider the inability of accountability mechanisms to generate “good enough” choices to engage and reflect citizens’ political commitments. Thus, Pitkin does not adequately tie accountability to the proper activity of representatives.

Pitkin herself provides a partial explanation for why her lack of attention to accountability is a problem. In her discussion of the uneasy relationship between democracy and representation, Pitkin warned that efforts to democratize the representative system have consistently led to representation supplanting democracy instead of serving it. “Our governors have become a self-perpetuating elite that rules or rather, administers, passive or private masses of people. The representatives act not as agents of people but simply instead of them” (2004, 339). Being a substitute is insufficient for genuine representation, according to Pitkin. Drawing on Hannah Arendt, Pitkin argues that “genuine representation” is possible “where the centralized large scale, necessarily abstract representative system is based in a lively participatory, concrete direct democracy at the local level” (2004, 340). The lively local political life allows people to learn and shape the meaning of democratic citizenship so that individual self-interest becomes consistent with public interests.

Because of the importance that Pitkin places on self-understandings, she rejects thinking about the choice of political commitments (or of representatives) as “a choice between two foods” (1968, 211). Politics should not be conceived as simply a matter of determining which preference produces the most satisfaction. For citizens’ self-understandings of their political commitments and their political practices influence the amount of satisfaction they derive from political processes. As a result, one cannot calculate political priorities as easily as a shopping list.

Pitkin also rejects using competitive economic market terminology to understand political representation. Such a way of understanding representation is likely to treat political issues as “questions of knowledge, to which it is possible to find correct, objectively valid answers” (212). The representative becomes an expert and the opinions of constituents become “irrelevant.” So if we treat politics as merely a mathematical question concerning the aggregation of existing preferences or as a scientific question about efficient methods of preference satisfaction, Pitkin concludes that the democratic commitment to “counting noses in the constituency” will appear “foolish” (212). Democracy comes to be seen as an obstacle to good representation. Pitkin’s understanding of the paradox of representation means that being the most efficient or capable at satisfying preferences won’t necessarily improve the normative value of the representative’s actions. Such efficiencies might improve a representative’s capacity to act and thereby the legitimacy of her authority, but representation involves more than that.

Throughout her career, Hanna Pitkin was drawn to paradoxes. In contrast to those who approach a paradox in order to offer some way to reconcile contradictory meanings or avoid the appearance of inconsistency, Pitkin highlighted the paradoxical elements of our conceptual language. She insisted that the theoretical complexities and moral nuances of political life require representation to remain a puzzle. Filtering through the everyday usages of the concept, Pitkin also reveals how the meaning of representation changes with use. Pitkin breathes life into the conceptual meaning of representation.

Pitkin, though, sketches the semantic landscape of a concept without providing a normative map about how we should navigate that landscape. She rejects absolutist principles and perfectionism. Humans must create the new directions and practices of representation themselves. This approach to political theory is best reflected in her response to the question, “Can democracy be saved?”: “I am old; it is up to you” (2004, 342).

This approach to political theory allows for humans to disagree deeply about political questions and their answers. For Pitkin, there are multiple ways of representing and having a legitimate state. The point of political theory is not to reach consensus about these ideas but to foster mutual understanding. If we continue to rely on representative processes to settle disputes among citizens, The Concept of Representation will help us understand the limitations and problems with those processes. In particular, Pitkin emphasizes the problem with adopting a limited and partial view of the concept of representation.

Still, Pitkin’s approach to surveying the meaning of political concepts might give us pause: evaluations of representative processes are likely to diverge if society disagrees too drastically about what good representation, let alone good democratic representation, means. Pitkin’s desire to make interests, preferences, and processes adaptable to various circumstances opens up that possibility that normative understandings can be strengthened or weakened by political practices. When confronted with deep disagreements and mutual misunderstandings, Pitkin recommends asking more questions, such as whose interests and well-being are served by existing representative processes. Such questions will hopefully generate new representative practices and the conceptual language necessary to better negotiate political conflicts. In the end, Pitkin challenges us to adopt various ways of viewing our political concepts and to critically question the function of certain political concepts within its political environment.

Castiglione, D.   2012 . “ A New Agenda for Democratic Representation? ” Politics & Gender 8(4): 518.

Google Scholar

Celis, K. et al. 2014 . “ Constituting Women’s Interest through Representative Claims. ” Politics & Gender 10: 149.

Diamond, L.   2002 . “ Thinking about Hybrid Regimes. ” Journal of Democracy 13: 21.

Dovi, S.   2008 . The Good Representative (Vol. 2). Oxford: John Wiley & Sons.

Google Preview

Grant, R. and R. O. Keohane . 2005. “ Accountability and Abuses of Power in World Politics. ” American Political Science Review 99(February): 29.

Griffiths, A. P. and R. Wollheim . 1960 . “ How Can One Person Represent Another?. ” Aristotelian Society Supplementary Vol. XXXIV: 187.

He, B. and M. Warren , 2011 . “ Authoritarian Deliberation: The Deliberative Turn in Chinese Political Development. ” Perspectives on Politics 9(2): 269.

O’Donnell, G. A.   1998 . “ Horizontal Accountability in New Democracies. ” Journal of Democracy 9(3): 112–126.

Phillips, A.   2012 . “ Representation and Inclusion. ” Politics & Gender 8: 512–518.

Pitkin, H.   1965 . “ Obligation and Consent—I. ” American Political Science Review 59(04): 990-999.

Pitkin, H.   1966 . “ Obligation and Consent—II. ” American Political Science Review 60(01): 39-52.

Pitkin, H.   1967 . The Concept of Representation , Berkeley and Los Angeles: University of California Press.

Pitkin, H.   1968 . “Commentary: The Paradox of Representation.” In Nomos X Representation, edited by J. Roland Pennock and John Chapman , 38-42. New York: Atherton Press.

Pitkin, H.   2004 . “ Representation and Democracy: Uneasy Alliance. ” Scandinavian Political Studies 27(3): 335.

Plotke, D.   1997 . “ Representation is Democracy. ” Constellations 4(1): 19.

Rehfeld, A.   2006 . “ Towards a General Theory of Political Representation. ” Journal of Politics 68: 1.

Rubenstein, J.   2007 . “ Accountability in an Unequal World. ” Journal of Politics 69(3): 616.

Saward, M.   2006 . “ The Representative Claim. ” Contemporary Political Theory 5(3): 297.

Saward, M.   2010 . The Representative Claim . Oxford: Oxford University Press.

Warren, M. and D. Castiglione . 2004 . “ The Transformation of Democratic Representation. ” Democracy and Society 2(I): 5, 20.

Zakaria, F.   1997 . “ The Rise of Illiberal Democracy. ” Foreign Affairs 76(6): 22.

For one answer to this question, see Dovi 2008 .

• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

14 stellar photos from the 2024 total solar eclipse

April 8th's total solar eclipse began on the Pacific coast of Mexico and ended off the Atlantic coast of Canada.

By PopSci Staff | Published Apr 8, 2024 4:35 PM EDT

Today was one for the history books as a total solar eclipse crossed North America. The sky first darkened in Mazatlán, Mexico on the country’s Pacific Coast. Torreón, Mexico saw the longest totality at 4 minutes and 28 seconds . It then entered the United States through Texas and traveled through Oklahoma, Arkansas, Missouri, Illinois, Kentucky, Indiana, Ohio, Pennsylvania, New York, Vermont, New Hampshire, and Maine. It entered Canada via Southern Ontario, and continued through Quebec, New Brunswick, Prince Edward Island, and Nova Scotia. The eclipse left the continental North America on the Atlantic coast of Newfoundland, Canada, at 5:16 p.m. NDT. 

Here’s how the eclipse looked at various locations, from Mexico to Canada .

And if you’re wondering what the eclipse looked like from space , NASA shared the view from the International Space Station.

Ever seen a total solar #eclipse from space? Here is our astronauts' view from the @Space_Station pic.twitter.com/2VrZ3Y1Fqz — NASA (@NASA) April 8, 2024

If you can, consider recycling or donating any used eclipse glasses. Visit Astronomers Without Borders to learn more about how you can recycle your glasses. If you are located in the path of totality, many libraries will also offer convenient eclipse glasses recycling locations . 

Like science, tech, and DIY projects?

Sign up to receive Popular Science's emails and get the highlights.

Help | Advanced Search

Computer Science > Computer Vision and Pattern Recognition

Title: learning state-invariant representations of objects from image collections with state, pose, and viewpoint changes.

Abstract: We add one more invariance - state invariance - to the more commonly used other invariances for learning object representations for recognition and retrieval. By state invariance, we mean robust with respect to changes in the structural form of the object, such as when an umbrella is folded, or when an item of clothing is tossed on the floor. Since humans generally have no difficulty in recognizing objects despite such state changes, we are naturally faced with the question of whether it is possible to devise a neural architecture with similar abilities. To that end, we present a novel dataset, ObjectsWithStateChange, that captures state and pose variations in the object images recorded from arbitrary viewpoints. We believe that this dataset will facilitate research in fine-grained object recognition and retrieval of objects that are capable of state changes. The goal of such research would be to train models capable of generating object embeddings that remain invariant to state changes while also staying invariant to transformations induced by changes in viewpoint, pose, illumination, etc. To demonstrate the usefulness of the ObjectsWithStateChange dataset, we also propose a curriculum learning strategy that uses the similarity relationships in the learned embedding space after each epoch to guide the training process. The model learns discriminative features by comparing visually similar objects within and across different categories, encouraging it to differentiate between objects that may be challenging to distinguish due to changes in their state. We believe that this strategy enhances the model's ability to capture discriminative features for fine-grained tasks that may involve objects with state changes, leading to performance improvements on object-level tasks not only on our new dataset, but also on two other challenging multi-view datasets such as ModelNet40 and ObjectPI.

Submission history

Access paper:.

• HTML (experimental)
• Other Formats

References & Citations

• Google Scholar
• Semantic Scholar

BibTeX formatted citation

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

• Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .