Categorisation, Categories and Prototypes
By Paweł Zieliński
This article is about categorisation as understood by Cognitive Linguistics. Before cognitive sciences started discussing the idea of how humans categorise different entities, elements of any given category were based on Aristotelian logic. That is a bird could be described as [+ feathers] [+lays eggs] [+flies]. The boundaries of each category were fixed and clear, with no 'gray area' between them. Each member had a list of sufficient points that it needed to possess in order to be considered a member of that category. That is how scientific categorisation worked for a long time: there were categories, each self-contained, which helped in the taxonomy. However, this was not considered a sufficiently good model of categorisation, because there are many different types of birds which either cannot fly, or their plumage is not on par with the rest of the bird family. That is why scholars like Eleanor Rosch et al. put forward an interesting theory about prototypes.
To introduce this theory I would like you, as the reader, to produce a piece of paper and a pen. Now, I want you to imagine a piece of furniture, and write it down. It is highly probable that you have written a chair, or a table. It is the most prototypical piece of furniture, that is why most people produce this answer. A chair or a table is more likely to appear in our minds, when prompted about furniture than , for example, a filing cabinet.
We have established that a chair or a table is a prototype. However, what exactly is this prototype. To define it in the simplest of ways, it is an example of a family, which first comes into our minds. Furthermore, by saying that an example is highly probable to occur also means that it is the least controversial. Of course, as is the case with most linguistic theories, prototypes are based on culture and relevant only in that culture. So the FURNITURE example would have produced a different result in e.g. Japan, where chairs are not as culturally relevant as in Europe, or America. Besides this, people in all languages possess the ability to categorise according to prototypes.
In order to add more complexity to the above theory, there were other terms created i.e. SUPERORDINATE, BASIC LEVEL and SUBORDINATE categories. The middle group has been already discussed in this article. It is the one which prompts the most common answer. However, what needs to be discussed are the other two categories. To follow the example already cited, FURNITURE is a SUPERORDINATE category, in which all the things like chairs, tables, cabinets, closets etc are included. One can imagine this as an umbrella covering all the individual pieces of furniture. A SUBORDINATE category lies below the BASIC LEVEL one. It is more specific e.g. a kitchen chair is an example of a SUBORDINATE category. Thus, the categorisation of the above example would look like this: furniture ---> chair ---> kitchen chair. The middle categories, or the BASIC LEVEL ones are the most likely to be uttered, SUBORDINATE categories are considered only after more specific information is required.
Let mi provide a different example to illustrate a different idea of prototypes and categories. In similar vein to the previous example, imagine a typical bird. What kind of characteristics does it have? In all probability it has got winds, it can fly, it can sing, lays eggs and can fly. A robin is a good example. What of a penguin? Is a penguin a bird? It is, however it does not come first to mind when prompted to name a bird, it also does not have most of the characteristics of a typical bird. Compare it with a robin, or a peacock. Penguins cannot fly, have a different type of plumage, cannot sing very well etc. yet they are considered a bird. Another example comes from Wittgenstein's Philosophical Investigations (1953). In the book Wittgenstein asks whether there is a common definition for all types of games. There is not, however chess, football, ice hockey, albeit very different, can all be considered a game. Why does this happen? Because of a notion of FAMILY RESEMBLANCE. In the penguin example, it does share some characteristics with other members of the category BIRD, which in turn share other characteristics with different members. Likewise with games, chess has got something in common with other games. That is why, with many 'branches' we can connect all members of a given category with each other. Some of those nods will overlap, some will not, but thanks to this a penguin can be linked to a peacock, or a robin, just like chess can be linked to bridge, or football.
As mentioned in the first paragraph of this article, before Rosch et al. started their work with categorisation, the notion had been that all categories have clear cut boundaries. When the idea of prototypes arose, it became clear to linguists and other scientists that category boundaries are not clear but fuzzy. There are no clear distinctions between what is a prototype and what is not. There were good and bad examples, of course, but there were no wrong examples. Imagine Mt Everest, it is a mountain, but when does a hill end, and a mountain begin? Or where does a mountain begin? All these boundaries are man made, but they are discussed. There are some arbitrary measurements for a hill or a mountain in geography, but from a layman's perspective, it is all fuzzy.