I was following a chain of references, in search of a possible dissertation topic, when I came upon the book “Creations of the Mind: Theories of Artifacts and their Representation” (2007), edited by Margolis and Laurence. I was really only after one chapter but I took my chances with a few of the others, just in case.
To my surprise the first part of the book comprised 5 chapters looking at the place of objects in philosophy. It turns out that there place is somewhat vexed and slippery. There is an inherent ambiguity to objects that philosophers find troubling. (I confess that I read the first chapter, ‘Social Ontology and the Philosophy of Society’ by John Searle, and then decided that was enough philosophy for me.)
As far as I can grasp, actually setting out a clearly delineated class of things called artefacts is difficult. An object may be made of many different objects that loose their identity when assembled but the assemblage loses its corporate identity when taken apart. The example of a bicycle was given. But there is also an ambiguity because the class of artefacts involves some difficult decisions about the role of human intentionality in their creation and their relationship to the human body. The latter point is interesting because it relates to some more commonplace ambiguities that surround what is (or is not) an object. Take, for instance, a small pottery house that I can hold in my hand; no problem here, it is an object and a piece of mass produced decorative art. Now think about a house that you might live in; isn’t that an object? Do we perceive, and think about, the house differently just because of its size? I’m conflating philosophy with psychology, which the authors are keen not to do, but it points to a full-blown debate that was found in the later sections of the book, which were more focused on psychological issues. Do we think about objects, or certain categories of objects, in particular ways because of the class that they are in? (More on this later.) And, how do we arrive at those classes?
The editors allowed the debate between the authors of different chapters to come out. In many ways, it was a familiar argument. Are we born with hard-wired modules for thinking about different kinds of objects in certain ways (Mahon & Caramazza, Chapter 10) or are we born with a powerful, domain general learning mechanism that allows infants to rapidly learn about the properties of objects in their world (Mandler, Chapter 11). Actually, Chapter 15 (by Hauser & Santos) proved to be an excellent guide to this debate, both in the way that they laid out the terms of the debate and in the way that they were open about their own position.
Hauser and Santos are particularly concerned with mental concepts of tools and seek to understand the way that such concepts function through behavioural experiments with non-human primates. Firstly, Hauser & Santos define mental concepts. Humans and other animals can attend to perceptual differences between objects and respond to them differently on that basis but that does not necessarily imply that they have a concept of an object. “Concepts are not mere collection of features, although featural distinctions certainly play a role. What makes the conceptual distinction more interesting […] is that particular [mental] tokens of a class are situated in relationship to other tokens and the organisation of tokens is mediated by a particular theory of how they cohere.” (p. 269). This theory of tokens requires categories to group objects within that they might be treated differently.
Hauser and Santos outline the positions in the debate about the origins of tool concepts in the following schematic:
They identify with Group 3 in this schematic and their reasons for doing so are interesting. Their work with tool-using and non-tool-using primates indicates that “when monkey and apes use an object as a tool they use different featural criteria than when they use an object as food or a landmark and [other non-human primates] recognise that shapes, size, material and orientation are relevant featural dimensions for a functional tool, while colour and texture are not. On contrast, [amongst other examples], when tamarins classify and object as a landmark, they use colour and shape as relevant dimensions but ignore orientation.” (p. 282) […] “At a general level, these findings lend support to the domain-specific view of concept acquisition. Even at very early ages and in the absence of task-relevant experience, non-human primates seem to parse the objects in their world into meaningful global categories – tools, food, landmarks and animals. Such evidence suggests that non-human primates may have innate biases to interpret their world in domain-specific ways.” So, on this work, primates “are biased to distinguish tool-like objects […] and these biases facilitate experience-based learning about different kinds.” (p. 283)
All well and good, this seems to resolve the debate. However, the book “Origins of Object Knowledge” (edited by Hood & Santos) was published in 2009 (i.e. 2 years after this book) and, in that book, the debate showed no signs of having abated.
Notwithstanding this apparent resolution, the editors included another chapter to frame the debate. In Chapter 6, “Artifact Categorization: The Good, the Bad and the Ugly”, Malt & Sloman make a very different point. They argue that the very notion of ‘categories’ of objects is deeply problematic and the experiments used to investigate how object categories function in human psychology are deeply flawed. Their review of experimental methods shows that those methods do not replicate the way that people make judgements about objects particularly well, thus the conclusions drawn from those experiments can be questioned. Actually, they end (p 123) by saying that “no coherent account of artefact categorisation is possible and ‘categorisation’ is not a coherent field of enquiry”!!
Malt & Sloman (p 121) state “that psychologists talk of artefact as coming in ‘kinds’, where kinds are taken to be stable, psychologically real groupings. [However, they argue] that not only are the psychologically meaningful groupings not stable, but there are no bounded groupings in conceptual space except by virtue of names associated with objects. These arguments suggest that the notion of psychologically real artefact kinds is not viable.” Instead, it is better to “consider them to be flexible and situation dependent. That is, each time mental activities take place that result in a grouping of artefacts, one could say that an artefact has been formed.” (p 122).
This bit really messed with my head! The other chapters that I read, whilst referring to each other, did not refer to this one, which I take to be an indication of how unpopular these ideas were. At the same time, the language of artefacts as flexible and situation-dependent groupings echoes the language of Self-Categorisation Theory, which states that self-identity is neither fixed nor stable but emerges in every moment from context-dependent social judgements. The idea that we might deal with the identities of objects in the similar manner to the way that we deal with our own is quite appealing and has the virtue of parsimony. I haven’t had a chance to follow this lead and, given that this topic seems unlikely to lead to a feasible dissertation question, I might never do so but it is food for thought.