Today’s entry in the Science and Metaphysics event comes from James Trafford, who has published on Metzinger in Collapse, and is currently working on a book about what he calls ‘revisionary naturalism’.
James Trafford, Why Realists Should (Probably) Be Rationalists Too
1. What’s a Little Heresy Among Friends?
This is an incomplete collection of ideas that seem to be worth articulating, but as a legitimate position require far more defence than can be offered here. Also, I don’t intend to spell out what it means to be a realist or a rationalist, other than the intimations suggested below. I should say that, where I do not accept many traditional rationalist doctrines such as the appeal to intuition; the idea that a priori reasoning is indefeasible, and perhaps also infallible etc.; I want to suggest that one particular element of rationalism that has been almost completely lost in certain quarters somewhere between positivism and neo-empiricism, is absolutely key to understanding how philosophical realism could be a coherent position.
The heresy is this. After all that has been done in this area of philosophy to rid ourselves of the linguistic turn and its excrescences, I think that we should relook at its supposed master. In fact, I think that, although he has been bastardized as much as any thinker could be, there is a reading to be made that makes him almost essential to the realist project. Who is it? Frege.
2. Putnam’s Problem
Let us define metaphysical realism as follows:
- Realism: There is a world of objects and properties that is mind-independent. (Where, for any subject matter ß, mind-independent truths about ß are not true in virtue of facts about thinker’s minds (including facts about discourse and concepts).
Putnam, who was once a strong proponent of metaphysical realism, has since abandoned it on the basis of problems regarding what he terms “conceptual relativity”. The idea is that concepts, including those which are most fundamental to describing the ‘way things are’ have many different and contrary uses (e.g. Putnam 1987; 1988
The suggestion … is that what is (by commonsense standards) the same situation can be described in many different ways, depending on how we use the words. The situation does not itself legislate how words like “object,” “entity,” and “exist” must be used. What is wrong with the notion of objects existing “independently” of conceptual schemes is that there are no standards for the use of even the logical notions apart from conceptual choices. (Putnam 1988, p. 114)
At first glance, Putnam’s problem for the realist may seem relatively easy to defend against; perhaps it reads simply as a restatement of a simple correlationist argument. But, consider the reasoning in detail. When we look at the ways in which we determine meaning as a whole, we can take it to be something like a theory – sets of sentences or propositions that we accept as truths about the way things are. If this theory is consistent it will represent the world in a certain way with certain domains consisting of certain objects and so on. The terms involved in the theory will be assigned semantic values relative to the truth of the theory. All that has been determined, however, is that the theory is consistent, and this does not seem to suffice for its truth. There are, for example, competing theories, of which we may want to say that the semantic values which are assigned there are contrary to their correct meanings. At this point, Putnam poses the problem that the realist in not in a position to choose between competing consistent theories. That is to say, reality can be described in incompatible though equally true ways, and this, he suggests, implies that reality is indeterminate apart from the ways in which we choose to describe it.
One example Putnam uses to make the point is an imaginary scenario involving Carnap and a hypothetical Polish logician (e.g Horgan and Timmons). He imagines the Polish logician to advocate the idea of mereological sums (i.e. that the sum of any two things is itself an object) and Carnap to deny that such objects exist. They are faced with some marbles on a table, and both are asked to count the objects on the table. Carnap states that there are three objects on the table (M1; M2; M3). The Polish logician counts the objects and states that there are seven (M1; M2; M3; M1+M2; M1+M3; M2+M3; M1+M2+M3). This is, perhaps a slightly abstruse metaphysical issue, but we might equally imagine a ‘common-sense’ realist stating that there are three objects on the table, and a scientific reductionist stating that there are 1×1020 objects, since he counts only electrons as objects (or something like that anyway).
Putnam suggests that this scenario is one in which there are two incompatible, yet equally correct judgments about the number of objects on the table – there is a genuine conflict, but each is making a claim that, relative to the way in which they use the concept “object”, is true.
If two incompatible claims are equally correct, it is easy to see why Putnam thinks that metaphysical realism cannot be correct, since, even if humans cannot ever know the details, there ought to be some truth about the ‘way things are’ that is not determined by the way in which are conceptual schemes articulate them to be. The problem has wider ramifications – it could be extended, for example, to an argument between ‘folk-psychologist’ and an ‘eliminativist’ about beliefs, or to a phenomenologist and a Metzingerian about the existence of the self. And, perhaps most importantly, it threatens to undermine the idea of scientific progress. For example, Dalton’s definition of the “atom” in terms of indivisibility must belong to a different conceptual scheme to the contemporary definition of the “atom”, but the realist doesn’t have the resources (according to Putnam) to say that the contemporary definition came about through scientific progress through the elimination of falsehood and increase of truths about the way things are.
If ontology and truth are simply internal to conceptual schemes which serve certain purposes, then there is no ‘ready-made’ world which is structured independently of those practices. And, if the world does not come in categories and objects which are divided up in certain ways, then it makes no sense to say that science (or metaphysics) can be objective in the sense that it aims to describe such natural and mind-independent divisions. In this regard, there is also no reason to prefer the conceptual scheme of science to those which are more pragmatic and “intuitive” since neither could be more ‘in touch’ with the structure of reality.
3. Refusing Putnam’s Semantics
The first response to Putnam’s problem, I suggest, is that we should refuse the semantic account that underpins the argument that incompatible conceptual schemes can be equally true. This involves refusing a theory of meaning that has been popular since at least the positivists, which I will call the ‘modern’ conception of meaning. This conception emphasizes public language and conventional linguistic meaning, and particularly the idea that ordinary, actual, use is what determines meaning. On this view, the meaning of a term is given in its use, and is fixed by what a reflective speaker would articulate. To briefly explain this position, it begins with an idea about concepts (e.g. Harman 1973; Sellars 1974; Field 1977) where a concept is what is expressed by a term, and concepts compose thoughts, including the content of beliefs. Most importantly is that concepts can be possessed by thinkers – concepts allow those thinkers who possess them to think about certain entities and objects. So, we can speak about the conditions under which a thinker can possess a concept (a concept’s possession conditions; e.g. Peacocke 1992) in terms of the concepts conceptual role. Roughly, possessing a concept requires a thinker believing certain statements involving that concept or accepting certain inferences involving statements containing that concept. The conceptual role of a concept can be so defined, in terms of a specific set of beliefs or inferences involving that concept (though only a subset of core inferences may, strictly speaking, constitute conceptual role). The conceptual role determines both the conditions under which a thinker may be said to possess the concept, and also they determine the identity of that concept, so that two concepts will be identical only if they have the same conceptual role. It is easy to see why meaning is equated to use on this theory, since the conceptual role associated with a concept represent what a thinker shares with others who possess the concept, and the conceptual role also determines what that concept means.
Although this theory of meaning is often seen as a direct heir to Frege, his position is more complex, and it suggests a way out of Putnam’s problem. Let us call the kind of meaning expressed by the “modern conception” conventional linguistic meaning. Frege holds that, conventional meaning and concepts do not coincide. The conventional meaning of a term may be articulated by use, or perhaps more precisely, by the best reflective characterizations that have been agreed upon by the experts. But, for Frege, the sense of a concept is not determined by actual patterns of usage, but rather it is also determined by the ideals of rationality. Whilst many aspects of Frege’s theory are controversial and unacceptable for many, Tyler Burge has made this latter point far more clear.
Burge has famously put forward the notion of “anti-individualism” – the idea that the nature and individuation of certain mental states necessarily involves relations between the individuals in those states and aspects of an environment which is the subject matter of those states (e.g. Burge, 1979). In brief, the idea is that there are thinker-environment relations that play fundamental roles in determining concepts, and it is for this reason that a concept can be grounded in more than the conceptual role one associates with it. For example, a thinker can possess the concept “elm” and yet fail to have any substantive discriminatory ability when it comes to elm trees – the thinker may not be able to distinguish between beech and elm trees, say. The key point is that such a thinker’s conceptual role does not fix the extension of their concept elm. But, nonetheless, the thinker must possess the elm concept if they are able to have propositional attitudes with the content of that concept; a thinker cannot express the belief that <all elm’s are trees> without possessing the concept elm. Ordinary understanding sufficient for making certain inferences and entertaining certain thoughts, (what we would take to be ‘normally competent’) often includes mistaken associations, including those that may be considered central to guiding the application of one’s terms – that is to say, central to the supposed possession conditions on a concept.
The reason being that the correct application conditions for a concept are themselves dependent on the subject matter of the concept, as well as the linguistic community within which the term is being used. Concept possession plausibly involves some ability to apply a concept in certain contexts, and some minimal forms of reasoning with the concept, but these may often not be enough to identify the concept or fix its referent, nor must they be veridical. This is plausibly the case even for logical concepts: there may be no concepts that require complete mastery in order to be said to possess the concept.
There are good reasons, therefore, to distinguish between conventional meaning and the concept itself, with the suggestion that, for example, Dalton possessed the same concept “atom” that we do, since the reference of “atom” is grounded by relations of scientific discoveries to atoms. What Dalton did not possess, (and perhaps we also) was a complete understanding of the concept. In this regard, Frege invokes a typically Kantian image of “mist” and “fog”; where we do not fully understand a concept “its outlines are confused as if we saw it through a mist,” (Frege, Logik in der Mathematik); “in such a foggily blurred manner.” The key suggestion here is that, whilst a concept may appear to lack clear application (as Putnam’s problem elucidates) this is merely an appearance that can be put down to the fact that no-one has yet fully understood the concept: the problem does not lie with reality itself, but with our own limitations.
4. Back to Putnam’s Problem
If we briefly outline Putnam’s problem as follows (cf. Sider):
- There are different incompatible meanings for a term T
- None of these fits the use better than the others
- Meaning is determined by use
- Conclusion: Therefore, T is indeterminate in meaning, and so there is no fact of the matter which of the corresponding theories is correct.
I have been attacking the 3rd premise by suggesting that meaning is not determined by use alone. Key to this idea is that we often have incomplete understanding of our concepts, which is fine for successful communication and ordinary thinking, but it is not enough to determine the truth-value of our terms. What is striking about Frege’s position is that even those most competent thinkers may not be in a position to articulate the correct, full understanding of concepts; that would require an idealization. Frege thus agrees with Kant’s idea that cognitive capacity follows the ability to make judgments since judgment is inextricable from theoretical activity. But, his understanding of concepts in terms of theorizing beyond any extant conceptual explication sets it apart from the ordinary notion of linguistic meaning.
The presupposition here is that our judgments are founded upon aspects of reality that no-one has yet understood, and so concepts can only be understood in terms of something like a Piercean ideal science. Unlike Pierce, though, Frege did not link ideal science to the idea of final agreement, since, bluntly, Frege thought that we might all agree on the wrong thing! It may be speculative, though I do not think overly so, to say that Frege is here concerned with an ideal with may be, in some sense, inhuman.
Far more needs to be said here, but they key point is that, according to Frege, there is no guarantee that human understanding would suffice for full understanding, even if we recognize the limitations on human agreement, perception and observation, capacities, and so on. This is because we should rather think of concepts as (ultimately) conditions upon truth which are to be understood in terms of a comprehensive true theory. This notion of truth provides for thought what might be called a regulative ideal where full understanding is guaranteed by a complete theory (rather than, e.g. traditional rationalist notions of intuition). The idea may well be key to the position of metaphysical realism for the simple reason that full understanding of our concepts is not independent of knowledge of their subject matter –roughly, concepts depend for their individuation on the actual nature of the objects to which they refer. So, we can answer Putnam’s problem with the notion of incomplete understanding, which, itself involves the rationalist idea of full understanding as a regulative ideal guaranteed by a comprehensive true theory about reality. And this is why realists should (probably) be rationalists too.