What is the current thinking on how the brain processes generalizations? For example, the abstract idea of a right angled triangle vs specific instances of right angled triangles? Or the general idea of a cat vs specific instance of cats?

How does a generalization live in the brain? What's the current thinking on how it is encoded?

Medieval philosophers point to such abstractions (that are not material timeless "things") to argue that something itself not material and timeless must exist to "sense" them i.e. a soul. Which is hogwash. But how does one state the nature of such things in terms of human cognition? What is going on when we consider such abstractions?

  • $\begingroup$ Is this section of a Wikipedia article close to what you're asking about? wikiwand.com/en/Concept#/Mental_representations $\endgroup$ – Seanny123 Jun 9 '17 at 14:19
  • $\begingroup$ Thanks! That's useful. Any pointers to how generalizations (and abstractions) actually live in the brain? Specific things are somewhat easier for me to understand - in the sense that they are encoded as visual memories. But how is a generalization encoded? $\endgroup$ – Ziffusion Jun 14 '17 at 2:18

The thing that comes to mind first are the Gestalt principles. Up front - the Gestalt theory of perception has been criticized as being a descriptive theory, and not providing much insight into the processes that lead to perception.

However, the principles do address your question, as the Gestalt theory dictates that humans visually and psychologically attempt to make order out of chaos, to create harmony or structure from seemingly disconnected bits of information. Much of the principles deal with figure/ground separation, but an important other aspect is similarity, and specifically

Gestalt theory states that things which share visual characteristics such as shape, size, color, texture, or value will be seen as belonging together in the viewer’s mind.

The principle of grouping (Quinn et al., 2002) is indeed fundamental here:

The similarity principle claims that elements tend to be integrated into groups if they are similar to each other.

- Quinn et al, Psych Sci (2002); 13(4)

  • $\begingroup$ Thanks! That is interesting. Other than the visual aspect of it though, how does an abstraction or generalization "live" in the brain? I mean, we generalize a whole lot of things that may or may not be visual, do we not? $\endgroup$ – Ziffusion Jun 11 '17 at 13:32
  • $\begingroup$ @Ziffusion, the question seemed visually oriented so I took on a visual approach to your question. The Gestalt principles go beyond the visual system though. $\endgroup$ – AliceD Jun 11 '17 at 16:17
  • $\begingroup$ I appreciate you editing the question to clarify it. But do you think that one can go a bit too far with the editing? It seems to me that the question is not exactly what I asked. It almost seems like you want to bend and twist it, till it fits some version of what you want it to be! :) $\endgroup$ – Ziffusion Jun 14 '17 at 1:22
  • $\begingroup$ BTW I get your point about how we tend to categorize things. But that's not exactly my question. What I am asking is more about how generalizations (abstractions) are encoded in the brain. Specific things, I suppose, would be encoded simply as visual memories (or some organization thereof). But how are generalizations encoded / handled by the brain. Thanks for all your help! $\endgroup$ – Ziffusion Jun 14 '17 at 2:22
  • $\begingroup$ @Ziffusion please roll the edits back if they are not to your liking. However, I certainly did not intend to twist it, only clarify. I just don't understand what a specific instance of cat is, among other unclarities, that's all. Clear questions tend to generate the best answers. $\endgroup$ – AliceD Jun 14 '17 at 5:56

Accroding to the excellent book The Big Book of Concepts by Gregory Murphy, in the prototype view every category is represented by a summary of representation features, with weights for each value of each dimension (e.g. color, hair length). When a new member of the category is encountered we add weights of the present features and subtract the weights of it's features that are not parts of a category, and if above a certain threshold then it is judged to be in category. The prototype is the generalisation that you are looking for.

This is an extract from the book The Big Book of Concepts by Gregory Murphy :

enter image description here


Perhaps you are interested in the study to analogy making. According to Wikipedia, an analogy:

[...] is a cognitive process of transferring information or meaning from a particular subject (the analogue or source) to another (the target).

An analogy is often considered between to somewhat seemingly unrelated subjects (e.g. an atom vs the solar system), sharing an abstraction, i.e. a general abstract concept. This abstraction may be similar to your "general idea of a cat", which can be specified to specific kind of breeds.

A quick google search resulted in the following review (French, 2002) of computational models that try to explain analogies. In short, there are symbolic models, connectionist models and hybrid models that each use different approaches. The rather unsatisfactory conclusion of the paper is:

enter image description here


French, R. M. (2002). The computational modeling of analogy-making. Trends in cognitive Sciences, 6(5), 200-205.

  • $\begingroup$ I welcome any feedback on the down vote I received. Would be very glad to improve my answer, but I would need to know how $\endgroup$ – Robin Kramer Jun 15 '17 at 13:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.