The virtues and vices of categorisation

The act of categorisation is a fundamental cognitive process that is used to attach meaning to objects. As such, it forms the basis for daily interactions both social and introspective; social in the sense that stereotyping (a form of categorisation) affects not only our thoughts, but also our behaviour when interacting with others. This process is also introspective in that the act of categorising external objects influences how that information is internalised (correctly or incorrectly stored depending on the structure of the agent’s categorical schema).

Categorisation influences our perceptions of the world in a very marked way. The main advantage this function brings is that it makes generalisations possible and useful. Without categorisation, communicating thought processes and disseminating information from our world would become a very long-winded and convoluted process. The versatility of grouping commonly featured objects together allows us to talk about things informatively while leaving out all the tedious descriptive stuff. Categorisation is also one way of allowing meaning to be attached to objects, thoughts and feelings. For example, the emotion of feeling ‘sad’ includes a vast range of varying mental states all bubbling and boiling away in a sea of unpredictability. The overall result, of course, is easily identifiable to us as ‘someone of negative affect’, but how would we accomplish this feat if we did not have access to categorisation? We would surely be paralysed by the overwhelming variation that individual differences in the expression of sadness brings. One possible function of categorisation is to work cooperatively with the sensory regions of the brain to help provide an overall picture or concept to use in working memory space. Take face recognition for example, many hundreds of fluctuating variables (shape, position, features, colour etc) that somehow are compressed and averaged into something that is usable by the brain. The act of categorising the facial features into a coherent whole allows not only recognition, but the activation of memories, stereotypes, future planning and emotions (among other actions).

Delving deeper, I wonder whether it is possible to describe a thing without falling back to categorisation? It seems not, as the very act of describing something seems to presuppose, if not require, the existence of categorisation. This ‘reductionist’s nightmare’ becomes apparent with a simple mental simulation. Try now to describe a common everyday thing without referring to pre-established categories. Take a humble kitchen drinking glass for example. Straight away I have categorised the item; I could have been talking about any glass at all, however immediately I have succeeded in creating a mental image of a glass, which was then refined by the sub category of ‘drinking’. The first category, glass, could elicit countless mental images of everyday objects. Those images would cluster around some variation of the bell curve (although how do you arrange ideas and concepts either side of the most frequent and central idea as in a ‘normal’ bell curve?), with the frequencies of each item starting off low and graduating up to the most common. Most likely the majority of conjured mental images will correspond to some fuzzy approximation of the everyday drinking glass. Each image will vary from mind to mind, however the overall category is well defined and usable in terms of conveying ideas. The brain is ready to receive and store the incoming information under the category ‘glass’. Again we attempt to describe the glass without using categories; in this case we take the reductionist approach and peel back another layer of physical form. A possible avenue for this is to describe the molecular structure; X many billions of Silica molecules arranged in formations just so, composed in turn of X number of silicon atoms and X number of oxygen atoms…and so on. The problem here is that we are still referring to categories. We are using words such as ‘atoms’, ‘molecules’ and ‘oxygen’; all are categories of physical things that are inclusive of all those objects that make up the said category. They still succeed in conjuring up a generic icon in the mind’s eye.

Or a different approach could be taken, and instead of trying to explain the constituent components of the item in question, the utilities are proposed. Our old mate the drinking glass would thus be described in terms of its usefulness (holds liquids), its actions (constrains, lifted to the mouth, poured), its influences on our bodies (delivers nutrients via the mouth) and even the processes that went into constructing it (Sven from Ikea, cheap Chinese sweatshop). It soon becomes obvious that no matter how hard we try to avoid the use of categorisation, it forms the basis for our thought processes. Whether it be categories of sub-components, materials and atomic structure or categories of behaviour, actions or origin, placing everyday objects into generalised groups according to their features is what gives it meaning. Without categories not only would the (traditional) communication of ideas be difficult, even impossible, the very essence of the stuff around us would be meaningless. In short, without categoisation, the external world looses its meaning.

But what of the negative aspects of categorisation. Perhaps the most obvious is the potential for errors; that is, incorrectly categorising something to a pigeon hole that it shouldn’t belong in. Due to the fundamental (and often unconscious) manner in which catergorisation affects the entire thought process, an error at this foundation level can spell disaster for the entire system. Subjective ‘errors’ in the categorising process become most apparent in social situations. I believe this is due to a low level sub-routine that uses social interactions to make refinements to the overall system; by observing the responses of other agents (in the form of behaviours) to the behaviour of the self (once the category has been assigned and a response elicited) the sub-routine compares and contrasts how effective and accurate the assigned category is in relation to the categories of others. In this way our individual systems of categorisation are kept in-sync, thus preserving the collective sense of meaning and making communication possible. If this is unclear, take the following example. Bill is attempting to explain a novel object to Joe. Bill states the object is a ‘Kazoolagram’ but this contains no meaning to Joe at all; his categoorical ‘set’ is missing this category with its attached label of meaning. The object’s properties are then described, and Joe responds by suggesting similar sets from his repository of meaning; “Well is it anything like a Nincompoop?” Here Joe attempts to refine his mental schemas and grasps at existing examples to attach meaning to this unique object. The banter continues, with both participants gauging the accuracy of their categorisations through the behaviour of the other agent. Eventually they agree on the meaning of the object. This brings us to another question; is meaning emergent (ie greater than the sum of the parts) or simply a cobbled together collage of pre-existing mental representations (limited by the extent of the agents prior experiences)?

It seems as though the process of categorising can be influenced by the pre-existing content present within brains, especially past examples and experiences of events or objects similar to the one in question. Meaning seems to be both an emergent property and a combination of past experiences in that individually, the features of a category are useless, however together and in partnership with the agent’s existing knowledge (making the process faster if both agents have similar experiences) categories flourish into useful, meaningful tools for the processing and transmission of information.

The point of this article was to expose the extent of categorisation and provide the case for it’s existence as an everyday, fundamental cognitive process. Sure, categorisation has its weaknesses, but more than compensates for them with its strengths. Categorisation runs deeper than most would realise; potentially providing insight into the very way in which brains receive, process and store information. Perhaps a more accurate and efficient process may arise if humanity succeeds in modifying the essence of cognition towards better ways of classing objects and describing internal states. Maybe the direct transmission of meaning brain to brain will supersede categorisation and allow for instantaneous communication between agents.