When considering the question of what a concept is one is faced with a choice. A concept is to some degree a theoretical term that we use to explain, predict and control the behaviour of various different creatures. Now as a theoretical term, a scientist can pretty much define the term as they want. An explicit definition is to be justified by its explanatory and predictive utility within a scientific theory. Now some philosophers argue against this approach because it leaves obscure the relation between the ordinary language concept of ‘concept’, and any theoretical definition of the concept of ‘concept’. Now a scientist could reply to this that it is not his problem to explain the relation between his theoretical terms and the terms of ordinary language. His job is just to construct explanatory theories which help predict and control the phenomena under discussion, if a philosopher wants to worry about the relation between theoretical terms and ordinary language terms this is their prerogative; but it is not the job of scientist to worry about such issues. In fact Dan Dennett has argued that the job of the contemporary philosopher is to help people understand the relation between the scientific image and the manifest image.

In this short blog I won’t offer an explicit definition of what a concept is. I will instead discuss areas where people typically attribute concepts to others and move down the scale towards more ambiguous cases where it is unclear whether attributing concepts is justified. My central message will be that attributing concepts is not something that is done according to a clear cut criterion. There is no magical metaphysical cut off point where creatures suddenly become concept users. There are rather areas where we have less or more justification for attributing conceptual abilities to others.

One area where there is no disagreement on whether we are justified in attributing concepts to others is in ordinary mature language using humans. If a person has a language then everyone agrees that they also have concepts. The reason we feel justified in attributing concepts to people who have a language, is because through tracking their linguistic behaviour, it is possible for us to determine their conceptual abilities. Thus if we have a person who is speaking about Dogs we can determine what their concept of a Dog is by questioning them. The person can answer a variety of different questions about their beliefs on the nature of dogs and from this we can rationally reconstruct what their implicit concept of a Dog is. The fact that a subject can provide justification for his beliefs about Dog’s gives us clear evidence about the nature of their concept of a Dog. This ability to justify and respond to questions by giving reasons is a more rigid criterion than merely being able to discriminatively distinguish between objects in one’s environment. As people like Bob Brandom have noted, if we used the ability to discriminatively distinguish between objects in our environment as our criterion for attributing concepts to others, then we would be forced to attribute concepts to rusting metal, thermostats, flowers etc. So to avoid this overly broad criterion where we are forced to attribute concepts to any entity no matter how dumb, some philosophers have argued that we should limit the attributing of concepts those creatures that have a language.

This move of limiting conceptual abilities to language users (hence to human beings) is typically justified by an appeal to the philosophical views of Ludwig Wittgenstein. This appeal to Wittgenstein’s philosophy as providing a clear dividing line between which creatures have concepts and which creatures do not is ironic. It is ironic because it portrays Wittgenstein as a kind of Platonist who is providing necessary and sufficient conditions for something to count as being a particular concept, and of demonstrating that only language users have the essential criterion necessary for one to be justified in saying they are concept users. This quasi Platonist approach is the opposite of the approach adopted by Wittgenstein so it is very ironic that his name is so often attached to this Platonist world view.

One of the many novel aspects of Wittgenstein’s ‘Philosophical Investigations’ is that he attacked the idea that concepts are the type of thing which have essential features:

“Consider, for example, the activities that we call “games”. I mean board-games, card-games, ball-games, athletic games, and so on. What is common to them all?- Don’t say: “They must have something in common, or they would not be called ‘games’” – but look and see whether there is anything common to all. – For if you look at them, you won’t see something that is common to all, but similarities, affinities, and a whole series of them at that. To repeat: don’t think, but look!-Look, for example, at board-games, with their various affinities. Now pass to card-games; here you will find many correspondences with the first group, but many common features drop out, and others appear. Now we pass next to ball games, much that is common is retained, but much is lost-Are they all entertaining? Compare chess with noughts and crosses. Or is there always winning and losing, or competition between players? Think of patience. In ball-games, there is winning and losing; but when a child throws his ball at the wall and catches it again, this feature has disappeared. Look at the parts played by skill and luck, and at the difference between skill in chess and skill in tennis. Think now of singing and dancing games; here we have an element of entertainment, but how many other characteristic features have disappeared! And we can go through the many, many other groups of games in the same way, and can see how similarities crop up and disappear.

And the upshot of these considerations is: we see a complicated network of similarities overlapping and criss-crossing:  similarities in the large and in the small.” (Wittgenstein: Philosophical Investigations 67)

Wittgenstein’s conception of concepts interms of family resemblances has found some empirical support in the work of Eleanor Rosch (proto-type theory) (1970) and George Lakoff (1987). If we adopt this anti essentialist approach we are left with a view where the concept of ‘concept’ is itself a crisscrossing network of family resemblances with no clearly defined essential features located in Platonic heaven.

When we adopt this approach, we are stepping into a similar world as the world of biology, where we stopped asking silly essentialist questions like “when exactly does one species change into another species?” a long time ago. Likewise the question at what moment exactly do concepts appear on the scene makes little sense. This isn’t a matter of adopting a simple “yes” “no” approach but is a matter of degrees of use in adopting intentional level descriptions.[1]

To see the advantage of adopting this approach let us consider the approach we adopt to people with Schizophrenia and various forms of Aphasia. To Platonic essentialists having a concept is a clear cut thing. If you have no language you have no concepts. The cut off point is language and there is little more to be said on the issue. Unfortunately for these Platonists reality doesn’t fit into this neat little binary opposition. The world of biology is a messy one. And these Platonists philosophers need to wake up to the fact that reality doesn’t always conform to their simple a priori intuitions.

Let us consider a case of a language using adult who as a result of a stroke has a form Aphasia which results in the person no longer being able to understand or speak a language. Now imagine[2] that this adult still has the ability to find his way around his environment, he recognises his family members, and responds to them in a similar though diminished manner. Now for the Platonist once this person loses his linguistic abilities as a result of a stroke we are forced to say that the person no longer has any concepts. This approach though seems odd. Say (as is quiet common), such a person still likes the same food, the same music, recognises his wife, his children, family pet etc. This person according to the Platonic essentialist has no concept of his wife, or of his children etc. At best we can say that the person is a reliable discriminator of objects in his environment in the same way as a thermostat is.

The above approach is not logically contradictory. Nonetheless there seems little point in adopting it. The fact that the person makes similar categorical discriminations post stroke,  than they did pre-stroke, and that we can predict the persons behaviour by adopting the intentional stance and attributing conceptual abilities to them seems reason enough to do so. And there is little reason to deny concepts to people with aphasia in the same way we would with people with severe Alzheimer’s who near the end do not seem to have any concepts what so ever. In short there is little pragmatic reason for adopting the approach a person with aphasia, and even less theoretical gain in doing so. None of this is meant as a proof but is rather meant as a reminder of the way we typically apply these concepts.

For a further exercise in imagination stretching consider the case of a person with schizophrenia who is speaking in word-salad. Now there is a long tradition of a psychiatrists and psychoanalysts trying to interpret the language used by a person engaging in word-salad as a coherent story with meaningful concepts. In a famous case Daniel Paul Schreber wrote a book ‘Memoirs of My Mental Illness’ while undergoing a psychotic break. On a straight reading the book is senseless word salad which is intermingled with bits of sense here and there. Nonetheless people like Freud, Lacan, and Matte-Blanco (using the apparatus of psychoanalysis) have managed to give the book a coherent intentional level explanation of Schreber’s book. In more recent times the clinical psychologist Louis Sass has offered an intentional level explanation of Schreber’s book using different theoretical apparatus (phenomenological apparatus).

It should be noted that the intentional level explanations of Schreber’s work remains at the level of a description. And as far as I am aware we have no way of deciding which (if any) of the interpretations of Schreber’s work is the correct one. Furthermore we have no evidence that therapy using either psychoanalytic techniques or the phenomenological approaches above are in anyway successful in treating schizophrenia on its own. Typically the best approach to treating schizophrenia is to use a combination of medical treatment combined with some kind of talking therapy.

The reason I bring up schizophrenia is because it raises interesting issues for Platonic Essentialists about concepts. Does a person who is undergoing a psychotic break have concepts at the time of the psychotic break? Well if the person has a language then according to the Platonic Essentialist he has concepts. But what counts as having a language for the Platonist? Here I think their strict criterion falls by the way side. Language is not a single thing. It consists of multiple interacting components: Syntax, Semantics, Morphology, Phonology, Pragmatics etc. There are mountains of empirical studies into the deviance in syntax, semantics and pragmatics in the word salad type utterances of people with schizophrenia. One wonders which essential components must fail before the Platonist will decide that the person with schizophrenia is no longer a competent language user and hence that we are not at that time justified attributing concepts to him?

On the loose naturalistic approach that I am adopting (basically Dennett’s stance-stance approach) these strict yes, no questions can be dropped. We know that talking therapy[3] can be quiet effective when combined with drug treatment. This fact alone justifies us in adopting intentional level interpretations to people undergoing a psychotic break and treating them as concept users. Appeal to arcane philosophical intuitions about some monolithic entity called language and its importance do not decide the issue. Rather our ability to predict, and hence manage the illness is what matters.

Similar ambiguities arise for things like Williams Syndrome and Specific Language Impairment etc. The fact is that the world is a messy place and the narrow binary intuitions that Platonists appeal to are not very helpful in dealing with such issues. The loose pragmatic approach is the best way of dealing with the complexity of the empirical world and avoids the trivial simplicities of the narrow Platonism I have described.

A critic could argue that the loose pragmatic approach I am adopting here will still result in a kind of free for all where we end up attributing concepts to thermostats, rusting metals etc. I am not sure that this fear is entirely justified. I am not arguing for an anything must go approach to intentional ascriptions. In his (1991) paper ‘Real Patterns’ used the work of mathematician Gregory Chaitin to specify precisely when we have discovered a real pattern. Dennett glosses it as follows:

“Of course, there are bound to be other ways of describing the evident patterns in these frames, and some will be more efficient than others-in the precise sense of being systematically specifiable in fewer bits.” Any such description, if an improvement over the bit map, is the description of a real pattern in the data.” (Dennett: Real Patterns p.33)

Using Chaitin’s approach we can distinguish between purely pragmatic intentional ascriptions to things like thermostats and when we are actually picking out real patterns in the environment. So we avoid the problems which people like Brandom predict about being overly permissive in our attribution of concepts to others if we theorise that creatures that are non-linguistic have concepts.

Finally I have been quiet critical of people who argue that only people with a language have concepts. I have criticised their approach as being too essentialistic. However I think that their scepticism about attributing concepts to non-linguistic animals is very important. This scepticism helps by constantly calling into doubt the constant assumptions of those who are certain that some non-human animals have concepts. This process of criticism can only be a good thing and helps avoid complacent theorising unconstrained by the facts.

[1] To see the Intentional Stance approach in more detail see Dennett ‘The Intentional Stance’ (1987) and Dennett ‘Real Patterns’ (1991).

[2] This is an imaginary case but cases of people with Aphasia which obliterates their language production and comprehension but do not significantly affect a person’s general practical ability to negotiate his way around his social and physical environment are quiet common.

[3] Both Cognitive Behavioural Therapy and Psychodynamic Approaches show similar levels of success when combined with drug based interventions.