Thursday, 9 August 2018

Some speculation about Piaget on schemata and Wittgenstein

One of the better aspects of academia is getting interesting emails asking questions. An ex student, John, asked me this the other day.

“Hi Tim,

I understand that Wittgenstein maintains the correct use of a word will always presuppose our ability to use it. i.e. there is no separation between concept and ability to apply it. This made me think about Jean Piaget’s ‘schema theory’ involved in child development. As infants learn to interact with their environments, they develop mental schemata which assimilate objects from the environment. A schema might be learning to grasp and, by assimilating a toy into this schema, the child is able to pick it up. Piaget was confident that the child repeats and practices schemata before their successful use in assimilation. Schemata which fail to assimilate in the intended way are subsequently revised. He also maintains that this way of learning is maintained throughout development and into maturity.

On Piaget’s account, language would classify as symbolic schemata, where a given word will have its schema as anticipatory knowledge of the external environment. Initially, this makes me think that Piaget advocates a method of empirical judgement that easily falls foul of [Wittgenstein’s] regression [of interpretations].

Trying to find compatibility between Wittgenstein and Piaget, I’d be tempted to suggest that a word-schema lacks meaning for the child (even though he/she may know the sound of it etc.) until it is correctly applied and accurately assimilates the intended aspect of experience. At this point, the correct use of the word will be simultaneous with its application, pace Wittgenstein.

Would Wittgenstein buy into this theory of learning, do you think? What status would he give to undeveloped word-schemata - perhaps as having the same standing as objective definitions i.e. without meaning until applied?


Sadly Piaget isn’t one of my authors so I spent an evening looking at ‘The Origins of Intelligence in Children’ to try the following quick, but possibly wholly misguided reply.

Schemata for Piaget start with examples such as grasping and develop into more concept-like instances. It is the latter that looks the better link to Wittgenstein but the former may shed light on our options. Let’s take the idea of grasping or ‘prehension’ (a word I seldom use, to be truth). Grasping seems to be something that floats free or successful and unsuccessful cases. That is, one might equally say of an action that it was an unsuccessful attempt at grasping or a successful grasping. Both have something in common. At some philosophical risk, we might use the same word (‘grasping’) for both. If so – and to put the point needlessly paradoxically! – grasping doesn’t reach as far as the world. I mean, there’s no necessary world-involvingness of any instance of ‘grasping’ (on this way of speaking). Grasping can be characterised in terms independent of the world as a free-standing action or event.

In fact, this seems wrong to me. Grasping is a teleological action and hence I think that the root concept is world-involving. If someone were grasping, then they were grasping something and we can ask what. They might try to grasp and fail, in which case we cannot ask what they were grasping with any guarantee of an answer (they may merely might succeed in grasping the wrong thing). What is world-independent and free-standing is the muscle and finger movements that coincide with successful and unsuccessful grasping alike. That inheres in the body and stops at the outer boundary of the skin. But for the moment, let’s loosely talk of ‘grasping’ to mean this and hence in accord with the earlier suggestion that the same word can be used for successful and unsuccessful cases.

Suppose, then, that we imagine a more concept-like schema on the basis of grasping. It too will be world-independent and will be applicable to orientations towards the world of a subject (necessarily an older child) where things go well and badly, where instances fit and where they fail to fit such a concept-schema. It now makes sense to ask how well the schema fits the experiences or the bits of the world experienced by a subject. Just as a grasping action that fits a mug may not work so well with stemmed wine glass, so perhaps a concept-schema for dog fits some instances of dogs better than others. Since it seems right to say that there will be some instances that cause even an experienced language user pause, this picture seems plausible at first.

But what is the analogue of the world-independent grasping action – or the muscle and finger movements that coincide with successful and unsuccessful grasping alike – for a concept-schema? Think of any free-standing mental (or neurological) state and call that the schema, if you like. Because it is free-standing it can exist however it stands to the world so it can cope with good fit, poor fit and lack of fit.

The problem now becomes: what gives such a free-standing state any kind of fit with the world? One possibility: an interpretation. But now, what encodes the interpretation: another free-standing concept-schema?

The diagnosis is that starting with a free-standing state is fatal for an account of thought’s directedness. Hence if Piaget’s account of schemata assume that they are world-independent and brought in better and worse contact with the world then they might fit some basic actions (especially when described in non-teleogical terms) but they won’t fit concepts. It will be a bad account of the conceptual. By contrast, a concept, on a Wittgensteinian view, is all in the reach to the world. It’s not in the mind. Or: it’s no more in the mind than the relation of being on the mat is in the cat.