METAPHORS OF MENTAL MULTIPLICITY

© Ronald de Sous
Department of Philosophy
University of Toronto
Internet: sousa@chass.utoronto.ca

From  Blick und Blick in Spannungsfeld von Sehen, Metaphern un Verstehen: (Schriften der Académie du Midi, Bd. 3), herausgegeben von Tilman Borsche, Johann Kreuzer, Christian Strub.  München: Wilhelm  Fink Verlag, 1998

Ever since Plato’s tripartite model of the soul, metaphors of multiplicity have been applied to the human mind. In the last decade or so, three other avatars this idea of internal multiplicity have attracted much attention. They are Derek Parfit’s model of the person as a temporal multiplicity of successive selves (a refinement of David Hume’s) Daniel Dennett’s "multiple drafts" model of consciousness, a model that is intended to constitute a radical rejection of the "Cartesian" model of the mind in all its forms, and George Ainslie’s "picoeconomic" model, which construes individual choices as the outcome of economic interaction among separate "interests" within the person. In this paper, I shall examine these three models (Parfit, Dennett, Ainslie) with the following questions in mind: To what extent is each metaphor capable of changing the orientation of our thinking about the mind and self? Is it possible to separate the rhetorical force of a metaphorical model from its cognitive content? Might the rhetorical force carry unwanted implications which a commentary on the model must disavow? Could these be avoided by presenting the model in some non-metaphorical way? Is the notion of a model distinguishable from the notion of metaphor? Is mathematical modeling, in particular, a species of metaphor? I conclude that metaphors can be characterized as nonsense, but nonsense that comes in at least three grades: Metaphysics, or true (but inspirational) nonsense; fruitful nonsense or scientific modeling, and mathematical modeling, (a class which perhaps should not be called metaphorical at all), the literal application of an abstract model to a concrete phenomenon.

NOTE: sorry, NO FOOTNOTES  HERE.

ALSO, there are a few Chinese characters which will require a Chinese browser to be legible.  Ignore them if they'd be unintelligible to you anyway.
 

Recently I have been dipping my mental toes into Chinese language and philosophy. From the one and the other I derive opposite inspirations: some fairly common observations about the relation between Chinese philosophy and Western philosophy suggest a certain paradox of metaphor; but reflection on the constitution of Chinese written characters point towards a way this paradox can be evaded.

After some remarks about conceptual relativism, I will turn to the secret insight into the nature of metaphor afforded by the study of the Chinese writing system. My real topic, however, is mental multiplicity, a metaphor of which I shall discuss one ancient and three interestingly different modern examples. These will bring me back to a more judgmental, not to say censorious, conclusion about the uses of metaphor in philosophy and psychology.

1. The Alchemy of Metaphor

Chad Hansen, a leading student of Chinese Philosophy, tells us that at the root difference between our traditions is what separates individualism (Western) from holism (Chinese): "Chinese philosophy is nonindividualistic in the sense that it is more coherent to interpret it via a part-whole rather than a one-many contrast." (Hansen, 1985 p. 40) Now what exactly does this opposition mean? and why does it matter? With a little ingenuity, it seems that each side of the contrast might be easily enough translatable into the other. If a thing has parts, then it has many parts; conversely, if a set includes many individuals, then it can be thought of as having those individuals as parts. Hansen explicitly rejects the "Whorfian" intention of this doctrine: "This is not to be understood as an argument that language constrains thought" (ibid). Yet the claim is sometimes made that our systems or traditions of philosophy—Eastern and Western—are almost mutually unintelligible. And this, the familiar claim goes on, is due to their preference for one or the other of these apparently more or less intertranslatable metaphors.

The Paradox of Metaphor

But what, if any, is the content of such claims? I see two grounds for skepticism. One focuses on the claim of incommensurability; the other, more general, stems from the analogical element in any metaphor. These lead to two versions of a paradox of metaphor.

Suppose, first, that the claim about individuals and holism is one of incommensurability. Hansen doesn’t sound this buzzword, but he might well object to my purported "translation" of parts/whole to one/many, that the difference between the two seems to vanish precisely because it is what is lost in translation.

But now consider this dilemma: when the above argument is explained to me, either I understand it, or I don’t. If I don’t, then the explanation has been useless. But if I do, then the explanation is not just useless but ipso facto false. For the explanation you gave me was in English, (or in terms of concepts familiar to me from Western philosophy), and if it was intelligible in English then it must be false that what was being explained is something necessarily unintelligible in English. Such considerations have given rise to the formulation, notably by Quine (1960) and Davidson (1982) , of a "Principle of Charity", in the light of which it sometimes becomes puzzling that we can ever misunderstand one another. Obviously that goes too far: surely it must be true that we can fail to understand one another, even if we can’t tell both sides of the story in intelligible English.

This has a close analogue in the first paradox of metaphor. The skeleton of this argument goes like this (we shall put a little more flesh on it below): Either the content of a metaphor goes beyond the literal claim that it replaces, or it does not. If it does, it must be unintelligible. If it does not, it is redundant. In either case, the metaphor is useless.

This first version of the paradox concerns intelligibility, or meaning. The second version concerns truth. Let me approach it by sketching a more sweeping attack on the capacity of any model, metaphor, or analogy to constitute any kind of reason for believing anything at all.

The argument is straightforward, and presupposes a widely accepted analysis of metaphor. On this view, metaphors imply a certain kind of resemblance between disparate types of phenomena. If I note a resemblance between two classes of phenomena, that resemblance necessarily relates to a particular respect of the phenomenon in question. To argue from analogy is to assume that the two things in question resemble one another in other respects also. Any of those other respects, however, may be precisely ones in which the two things are disanalogous.

Note how this argument resembles and differs from Hume’s argument against induction. Hume (1975) says that it is question-begging to argue from the observed to the unobserved, on the basis of the assumption that the unobserved probably resembles the observed. Nevertheless, such inferences are precisely the kind we are programmed by nature to perform, and no amount of rational resistance will stop us.

Here’s how the present argument resembles Hume’s: The mind’s natural tendency to generalize is here also at work: something like induction is going on when we reason that if A and B are alike in n respects, they may well be analogous in the (n+1)nth.

Here’s what seems to be a difference: In the case of Humean induction, the inferred respect is precisely the respect in which we have actually observed the resemblance in the past. Metaphors, by contrast, implicitly claim that an object is likely to resemble some original in respect A because that object resembled the original in some other respect B.

But this difference is illusory, as Goodman (1983) showed. It assumes that time and place don’t alter cases. The similarity between past and future cases abstracts from the fact that all previous cases had a feature that no present or future ones have, namely being before now. So metaphors, considered as miniature arguments from analogy, can indeed be regarded as just particularly feeble inductions. And the skeptical case against them is only as good as the case against induction. That case is indeed unanswerable, as Hume claimed, but Hume’s deeper lesson is that it doesn’t matter because it’s based on asking the wrong question. The right question is not: how can we show that our inductions are justified, but rather, what are the practices that we actually engage in?

If this is right, the a priori arguments against the metaphorical enlargement of our understanding fail. But that is not to say that all arguments of this form are good: on the contrary, it is to insist that they have to be looked at individually. And that leaves us pretty much where we started.

Or maybe even a little worse. For the injunction to look at cases individually is unhelpful. Looking is fine, but what are we to make of what we see? To make something of particulars, we need universals—or at least some general principles or practices.

Understanding these general principles would take us far towards the discovery of the "laws of the mind", or at least towards a solution of the psychological problem of induction, by means of an understanding of the psychological basis of metaphor as the root of all thought. My ambition is not so overweening as to reach for these answers. Nor do I want to go further into the question of whether dualism or polarity represent a real or a spurious opposition. Rather I want to take a detour, and tell you a little, not about Chinese philosophy, but about Chinese characters.

The Creation of Sense

Chinese characters are a wonderful metaphor for metaphor. While the origin of characters is no less mysterious than the taxonomy of philosophical assumptions, they constitute a paradigm case of the creative work of metaphors, of their power to create a truly "emergent" phenomenon. ‘Emergence’ here refers to the fact that the meaning of a character often could not have been predicted on the basis of its parts and their arrangement. Characters work like molecules of sense. They are (in the main) made up of atoms of sense: but the properties of the molecules are not predictable on the basis of even the fullest knowledge of the phenomenal properties of the atoms. What we can observe about oxygen and about hydrogen will tell us very little about what water is like. And yet, undeniably, oxygen and hydrogen is what water is: in some subterranean, inaccessible way, the nature of water is brought to us by the nature of its components. Yet they are changed, "changed utterly". So it is with Chinese words. A few examples:

First, to bring out the contrast, an example of what I’m not talking about. Some characters are literal to the point of being pedestrian. The character Àá leì, for example, means "tears": and what could be more obvious than the juxtaposition of Ë® shuï "water" (in its "three-dot" form) and Ä¿ mù "eye"? The simplicity of this character, uncharacteristic as it is, may express the literal minded spirit of the proletarian revolution: for it is the "simplified" version of a character that at least a little more subtle, joining as it does the water character with another which expresses a more abstract kind of evil, ìå lìand is itself made up of a dog È®

quän under a window hù »§ .Contrast ο wei comfort: a corpse ʬ shi, expression, ʾ shì, an inch ´ç cún, topping a heart ÐÄ xin. Clues, perhaps, but leading only with some indirectness into the meaning they have come to express.

Consider now the combination of ÊÏ shì "family or clan" sitting on top of ÈÕ rì "day, or sun"; these together make up a character hùn, that means "dark, confused". When the character for woman Å® nü is added on the left, you get »é hùn, "marriage". I doubt, however, that any Chinese scribe ever brought to mind the awful appropriateness of the resulting message: that marriage, for a woman, represents the shadow of the clan darkening the sun of all her days.

One more example. ÏÊ xian, ¡°fresh": the juxtaposition of Óã yú, "fish" and Ñò yáng, "sheep". Why? "because fresh is what both fish and goat [meat] need to be"? I’m not sure if this is an example of any kind of trope that has a name. (Would it be a catachresis to call this a catachresis? Or perhaps it’s a strange sort of metonymy.) But it’s a way for a meaning to be generated out of previous meanings, an example of how, when meanings are juxtaposed, the resulting molecule is one the properties of which could not safely have been predicted. You need to tell a story to make sense of how the atoms combine to produce something with the properties of the molecule. But even then, the story doesn’t always help.

Chinese characters dramatize the point in a picturesque way. Actually, though, much the same can be done with English or French: e.g. why do the concepts of under, standing, together yield understanding? Or why does between and holding (entre, and prise), make entreprise? Why is running in not the opposite of running out, nor running up the opposite of running down? why is feeling run down to being run down not as feeling is to being? and so on.

These examples are supposed to remind us of the fact that some of the ways in which metaphor works are mysterious, subject at best to classification but possibly not to codification. Obviously they are all that. At the same time, they manage to be effective, effective in creating real meanings which make a difference to the repertoire of thoughts available to us.

I can now approach metaphors of mental multiplicity with the following specific questions: are such metaphors nothing more than question-begging arguments from analogy? Or do they pack the power of emergent molecules of sense such as are evoked by Chinese characters?

Gestalt, Visual Aid, or Emergent?

My thesis, in a nutshell, is that the content of metaphors and models in philosophy of mind need bear no relation to their vividness; indeed, the most striking metaphors may have virtually no content. I hope to make a start on the question of what makes the difference between models that do and those that don’t make a difference to the content of our understanding. But what is it to make such a difference?

What I called the metaphor for metaphor which I sampled a moment ago has both a positive and a negative aspect.

The "molecules of meaning" that emerge from combinations of semantic elements are at least sometimes the results of mechanisms that are effective, in the sense that they create real meanings which make a difference to the repertoire of thoughts available to us. That’s the positive aspect. I called these Chinese characters metaphors for metaphor, because the alchemy that governs their composition is mysterious, difficult to fathom, recalcitrant to the application of simple inductive rules. On the other hand, they actually work this way, and though some might have been manufactured with just this or some other idea in mind, most of them are the result of spontaneous synergies of sense.

But this is also what generates the negative aspect: if there is no systematic way these molecules are built up out of their atoms, does this not mean that it is futile to pursue the mechanisms responsible for the results of their combination? That indeed would be resorting to "individual cases" with a vengeance. The problem of commensurability and the parallel with the problem of induction have deep implications for the role of models in science. The moral is not just that a metaphor’s impact is sometimes merely illusory, but that we might have no systematic way of finding out when it is and when it isn’t.

To give the feel of this idea, allow me to remind you of a rather well-known story.

Somebody remarked to Wittgenstein that it might have been reasonable to believe that the sun revolves round the earth, because it looks as if the Sun were revolving around the Earth. Wittgenstein’s reply was to ask: And what would it look like, if the Earth were rotating around its axis while revolving around the Sun?

The influence that a picture can have on us can be misleading, not because the picture is misleading in itself, but because it misleadingly suggests that there is a difference that this particular picture can make.

In the remainder of this paper, I want to illustrate this idea in terms of some of the models of multiplicity that have been applied to the human mind. The sketches I shall offer are brief and coarse; I certainly allow that zooming into any of them in more detail might alter the picture entirely; but that is a risk I must take.

2. Models of Mental Multiplicity

Of my four examples, some more than others represent a change in our concept of the mind; some more than others fit readily into place among our existing common-sense concepts; some more than others represent blueprints for research into mechanisms of which they constitute explanation-sketches. All raise questions about how metaphors and models function to promote real or specious understanding.

My first example is also the first developed psychological model of the mind as multiple. It is Plato’s tripartite soul in the fourth book of the Republic.

Plato’s model is motivated by the need to explain inner conflict. Among its most intriguing features is the fact that, like Freud’s (1933) model of ego, id and superego, it involves three rather than two parts, incorporating a correspondingly more tricky argument for the separation of each part from both of the others. But here I only propose to look at the metaphor of conflict as inconsistency which underlies the proof in question. I believe it holds lessons for modern psychology and philosophy of mind.
In brief, the argument goes like this:

Thus paraphrased, Plato’s argument is open to the first paradox of metaphor as sketched above:

Either the content of the claim that the mind is multiple does not go beyond the content of the claim that there is inner conflict, or it does.

In the first case, we may still feel that the metaphor has changed our view of mind and will. But other than having a certain picture, we may be hard pressed to say what has really changed. Paraphrasing Wittgenstein’s boutade, one could here ask: What would the mind have been like if it hadn’t been divided "literally", but only "metaphorically"?

Taking this question seriously would require us to provide additional content to the multiplicity of the mind. For metaphorical multiplicity, the experience of inner conflict suffices. But that is the premise of the analogy, not its conclusion. An inference to literal multiplicity goes beyond the evidence contained in that premise. The contradiction on which that inference rested was, in fact, only metaphorical. So the argument might just as easily be taken as a reductio of the claim that the multiplicity of the soul follows from the metaphorical contradiction exemplified in the experience of conflict. A clear apprehension of this reductio might have prefigured the momentous discovery that intentional relations do not obey the laws to which real relations are subject.

Thus a critical look at Plato’s argument might have led to the discovery of the peculiar properties of intentionality. In itself, however, the metaphor of inner conflict doesn’t do any more than describe experience, for the feature of multiplicity that appear to be explanatory has no independent and non-question-begging existence.

In sum, Plato’s model contains three lessons. First, in its strict failure to establish its intended conclusion, it exhibits the paradox of metaphor (in its second form): for it fallaciously infers a real existence of multiplicity in the soul from the existence of metaphorical multiplicity in experience. Second, in its evocative power, it transcends the paradox by forcing us to think of the mind as multiple, even if the empirical meaning of this claim remains obscure. And, third, its very failure holds the potential to create meaning, by forcing recognition of a defining feature of intentionality.

3. The Fictional Continuity of the Self.

 My second example is Derek Parfit’s widely discussed model of the self as having only fictional unity through time. I submit that although this metaphor has generated a large amount of heat in the philosophical community, it is actually wholly devoid of content. It’s as if the model transcended the paradox of metaphor without actually facing it in the first place; for, as we will see, there is a suggestiveness about it that is not traceable to any real consequence that adopting it might have.

The Hume/Parfit model is not meant to address the issue of conflict. The element of multiplicity in the model it presents is multiplicity through time. Hume and Parfit both begin with the idea that strict identity (of a sort that obeys so-called Leibniz’s law) is not to be found in objects that persist through time, since it cannot be said that everything about a person S at time t is true of S at time t’. From this premise, they infer that strict identity cannot really be in question when we debate whether this of that represents a criterion of identity through time, or a criterion of reidentification. They attempt to reconstruct what is actually at stake when such issues are raised, and draw our attention to what we care about in such cases. Hume concludes that our sense of identity through time is a mere illusion. We could think of it as like the apparent motion in a movie: on the screen, nothing moves: we know this yet cannot help ourselves seeing what is on the screen as moving.

On the most straightforward interpretation, however, this theory is half-baked. For rather than concluding that identity through time is an illusion, Hume could have claimed that the conditions he identifies at the root of the illusion are actually defining conditions of what we call personal identity through time. This, roughly, is the twist put on the view by Parfit, who seems to agree that strict identity through time is a contradiction in terms, but insists that this is not what matters anyway. What matters is rather the "R-relation" of connectedness (through memory) and continuity (through overlapping threads of connectedness).

This theory is meant to be a guide to our thinking about the self: about problems of distribution, of justice, and of responsibility. The consequences claimed for it are far reaching. None of them follows, however. Rather, the alleged consequences are merely things we are inclined to think in the light of a certain picture.

One alleged consequence of Parfit’s view, for example, is that once the "important" relation of R-relatedness has been substituted for the old relation of identity, we must countenance a number of possibilities that from the old perspective seem absurd. These possibilities are generated by certain thought experiments. One of these involves the transplant of the two halves of a split brain into different bodies. Since both would be `R-related’ to the original person, both resulting persons would be (in common parlance) said to be "identical" with that original. But the ordinary concepts of identity and non-identity are both transitive. And in the case of this sort of "fission", A=B, A=C, but B C—which is absurd.

But is this really so? Parfit can resist the unwanted inference as follows. I said a moment ago that Hume’s conceptual innovation is in a certain sense half-baked. What I mean by that is that a further conceptual innovation will in effect cancel out the strangeness introduced by the first. And so it goes here too: we can translate the claim that B is identical to A as the claim that both A and B are part of a single spatio-temporal worm, AB. Similarly, A and C are both temporal parts of AC. So we don’t have to fear that we shall have to admit identity between C and B: they are merely different parts of two different things which actually happen to share a part. And there’s nothing offensive to common sense in the idea that two things might share a part.

One of Parfit’s arguments in favour of this view is that it offers certain philosophical consolations: ways of seeing things that result in rather different feelings that one might now have about what it is like to be human:

In short, Parfit’s view of the self, no more and no less than the view he opposes, are metaphysical is just the pejorative sense given to that expression by the logical positivists. They have no literal content and make literally no sense. Eppur, one might exclaim, ci muove! And yet, it is a view that has the power to move us. We feel that somehow it means something and makes a difference. It’s just that whenever we try to state what difference it makes, the difference fades away.

4. Dennett’s Multiple Drafts Model

The main thrust of Dennett’s "multiple drafts model" of consciousness is a negative one. His aim is to undermine the prevalent picture of the unified self, epitomized by his image of the "Cartesian Theater" in which a single observer, consciousness, witnesses the passing show of a single life. His motive is a scientific one: the old picture should be replaced, because like Ptolemy’s astronomy resisting the Copernican it impedes the progress of our understanding. It forces us to construe certain phenomena in an artificial way, to fictionalize aspects of ourselves.

In this case, the metaphor being rejected is clearer than the one being advocated. It is clear enough, indeed, to be subjected to a decisive experiment. This is the result of the famous phi phenomenon—an experiment which would also have given comfort to Hume. (Dennett 1991, p. 120ff.)

When exposed to two bright spots lighting up in rapid succession in different parts of the visual field, the eye interprets them as a single moving dot. More impressive is what happens when the dots are different colours: it then turns out that the moving dot is seen as changing colour half way in between. Now on the model of the mind as a single theater, this requires us to choose among two more pictures, described by the metaphors "Orwellian" and "Stalinesque".

On the first, something happens—a revision of history—after the original experience and before it emerges as a memory to the theater of consciousness. On the second, we must posit something like an "editing room" in which the information is processed before it is allowed into consciousness. Like Plato’s conflicting selves, these two pictures conflict. But unlike Plato’s conflicting selves this cannot be dismissed as a harmless discovery of the peculiarities of mental theorizing. Valid explanations cannot conflict. But Dennett argues that no crucial experiment could possibly decide between the two pictures:

I have given only the briefest sketch of Dennett’s model. It is sufficient for my purpose here, which is merely to illustrate how the metaphor of multiplicity works, in this case, to guide scientific modeling. Unlike Parfit’s model, which I called "metaphysical", the scientific use of models isn’t really subject to the paradox of metaphor, and therefore need not transcend it. It does not infer unobserved analogies from observed ones. Instead it offers analogies for testing. The function of Dennett’s model is to push us in the direction of science. The fact that we more naturally call it a model than a metaphor may reflect the fact that metaphors proper demand both the paradox and its transcendence.

 5. Ainslie’s Picoeconomics.

George Ainslie’s (1992) book is subtitled "The Strategic Interaction of Successive Motivational States within the Person". It is an exploration of a number of psychological phenomena, many of them phenomena of pathology, in the light of a very simple question about the rate at which we discount the future.

"Discounting the future" is, I suppose, a metaphor: to be specific, it is a metaphor from economics. But it is one of those peculiar metaphors that do not appear to have any literal equivalents. Or rather, like Plato’s, it is meant, in some sense, to be literal: the "interests" that are at the root of it are supposed to be (to some extent) independent entities (with at least the same kind of logical independence as is featured in the Hume/Parfit view.

Ainslie’s theory purports to provide a unified explanation for a surprising variety of phenomena. These include (among other things) patterns of resistance or surrender to temptations, addictions, obsessive behaviour, procrastination, sexual exhibitionism, anorexia nervosa, Don Juanism, gambling, inconsistent preferences, rigidity of character, and more. This harvest is a rich one indeed; so rich that it might well lead us to harbor some suspicion about the empirical content of its claims. If we ask, what sorts of models are likely to apply to all sorts of different things, the obvious answer is: mathematical models. Indeed, if I am right about Ainslie’s theory, it is less a metaphor for mental multiplicity than an application to certain mental phenomena of a certain mathematical theory. It may, in this way, suggest that contrary to what we might have thought, mathematical models are not a species of metaphor. In fact, I shall argue, mathematical models are literal, and it is their application, if anything, that is metaphorical.

Before attempting to clarify these rather dark sayings, let me briefly explain the most paradoxical consequence of Ainslie’s theory of future discounting. This involves a reconceptualization of pain. A reconceptualization of the all-too-familiar phenomenon of pain in the light of a different model will raise several questions: To what extent can it make a difference, in the sense I am seeking to narrow down, to the way we think about pain? Is it nothing but a picture? Is it a good picture? In what sense?

The raw materials of the theory are these: (1) a model that explains why our rate of discounting of the value of future goals leads to temporary inversions of preferences; (2) a puzzle about the nature of pain; and (3) the application of 1 to the solution of 2.

(1). By analogy with distant objects of sight, future probabilities seem to diminish in size as they recede into the future. Given the uncertainty of life, any given prospect simply gets less probable as other possible events crowd in to possibly thwart it. Its net desirability can then be thought of as its absolute value weighted by its likelihood of coming about. This, in any case, is one way of accounting for the observed fact that people regularly discount the future.

But mere discounting, and the rationality of mere discounting, is not what is at issue. Rather, it is the shape of the discount curve that counts. Ainslie shows that if it is merely exponential, temporal perspective will not disturb the relative ranking of variously distant prospects. This means we won’t be able to explain why, as we get closer to some objectively less desirable goal, that lesser goal can seem momentarily more attractive. On the other hand, if the curves are hyperbolic, (growing to infinity—except for a constant in the denominator—when the delay goes to zero), the apparent desirability of the lesser goal can temporarily dominate the ("objectively") larger but more distant one. (Ainslie, 1992, pp.63-64). The effect can be seen intuitively, though less precisely, by attending to the masking effect of taller structures by smaller ones when one gets closer to both.

2. Now the puzzle about pain is this: we attend to pain, though there are many examples, such as of athletes unaware that they have been wounded in the game, to show that this is not a biological necessarily. Yet pain is obviously aversive.

We frequently "worry" our pains rather as we scratch our itches: there is this much phenomenological support for this idea. But apart from that, it has very little appeal as a description of the feeling of pain. Moreover it has the awkward consequence that it apparently assimilates feeling pain to being addicted. For addiction exhibits the same pattern: the pleasure dominates in the early phase, even though the total discomfort over time outweighs the sum of pleasure. This raises a startling question: "To be consistent, must not society either forgive all drinking or blame all victims of pain?" (op. cit., 113).

This looks like an instantiation of the paradox of metaphor: if pain is in this respect like addiction, it no more follows that it should be treated like addiction than it follows from the existence of inner conflict that the soul must have separate parts. But in fact it demands no transcendence. Economic models are mathematical models: inferences drawn from them are therefore deductive, or they are nothing. Mathematical models are, in a vacuous sense, true of everything. Does this make them trivial? No, for it’s an empirical question whether the model can be usefully applied in the given case. If the model applies to pain at all, then what it says of the process of pain is simply literally true: pain has the mathematical properties that follow. If those properties are interesting, then the application was useful. In any case, no other inferences are licensed, and mere connotations—such as might be suggested by the comparison with addiction—should be ignored.
 
6. Conclusion: Some Grades of Nonsense

Plato’s original model of mental multiplicity illustrated the paradox of metaphor, as well as the possibility of its transcendence in the alchemy of meaning creation. The more modern metaphors I looked at in this paper seem to divide naturally into three classes: True nonsense—which may yet prove inspirational (the metaphysical); fruitful nonsense (or scientific modeling); and a class which perhaps should not be called metaphorical at all, namely the literal application of an abstract model to a concrete phenomenon (mathematical modeling).

True inspirational nonsense was exemplified by the claims and counterclaims generated by Parfit’s views of the self. While they seem to create meaning, the meaning they create seems entirely metaphysical, in the sense of being devoid of empirical content.

Fruitful nonsense is exemplified by Dennett’s multiple-drafts model. The thrust of Dennett’s arguments is mainly negative, aiming to discredit certain pictures that have inspired philosophical psychology in the past. But every aspect of the analogies these pictures, and Dennett’s rival picture, suggest is immediately put up for some sort of testing. So these models raise no paradox of metaphor.

Finally, there is a kind of metaphor which puts up for testing not analogical inferences but mathematical consequences. They too, as illustrated by Ainslie’s initially implausible theory of pain, escape the paradox. Yet they too can retain a certain mysterious element of meaning creation, insofar as they demand that we see some familiar phenomenon in unfamiliar terms.

These reflections may throw some light on a curious puzzle about the history of recent philosophy: How exactly did the positivist critique of metaphysics fail? Many agree that it did fail, but the arguments are often unconvincing. After all, what could be wrong with the requirement that a statement is meaningful only if its truth or falsehood can make a difference to experience? One answer is that the requirement is often saddled with an inappropriately narrow concept of experience. But too inclusive a conception of experience also has its dangers, since it suggests that there will be nothing that the positivist critique can eliminate as nonsense after all: Ayer’s (1952) paradigm of rubbish from the Hegelian canon, for example, presumably is capable of engendering an Aha! experience in some people A slightly more restrictive criterion of what is allowed to count as experience might demand that any metaphysical claim come with a scientific research program of some sort. The seriousness of such programs would serve to differentiate different grades of metaphysics, merging into nonsense at one end of a continuum and into science at the other.

In the intermediate ranges, we should perhaps allow the alchemy of metaphor to work its own magic in the creation of emergent meaning, even where this is not accompanied by the sort of content that would meet positivistic strictures. If the stricter demands of positivism could always be met, we probably couldn’t hope to say anything interesting about metaphor. Indeed, we probably couldn’t talk at all.

Ayer, Alfred Jules. 1952 <1936>. Language, truth and logic. New York: Dover.

Black, Max. 1962 <1954>. "Metaphor. Proceedings of the Aristotelian Society 55. Reprinted in Models and metaphor. Ithaca: Cornell University Press.

Churchland, Paul M. 1984. Matter and consciousness. Cambridge, MA: MIT Press.

Davidson, Donald. 1982. Inquiries into truth and interpretation. Oxford: Oxford University Press, Clarendon.

de Sousa, Ronald. 1997. "What can’t we do with economics? Reflections on Ainslie’s Picoeconomics". Journal of Philosophical Research 22:197-209..

Dennett, Daniel C. 1991. Consciousness explained. Boston, Toronto, London: Little, Brown.

Freud, Sigmund. 1964 <1933>. New introductory lecture on psychoanalysis, Ed. and trans. James Strachey. In Standard edition of the complete psychological works of Sigmund Freud, 3-183. London: Hogarth Press.

Goodman, Nelson. 1983 <1955>. Fact, Fiction and Forecast. Cambridge, MA: Harvard University Press.

Hansen, Chad. 1985. "Individualism in Chinese thought". In Individualism and Holism: Studies in Confucian and Taoist values., pp. 35-57. Ann Arbor: Center for Chinese Studies, University of Michigan.

Hume, David. 1975 <1777>. Enquiry concerning human understanding. Ed & introd by L. A. Selby-Bigge. Revised by & notes by P. H. Nidditch. 3d ed. Oxford: Oxford University Press, Clarendon.

------. 1978 <1888>. A treatise of human nature. Ed. L. A. Selby-Bigge. Revised by & notes by P. H. Nidditch. 2d ed. Oxford: Oxford University Press, Clarendon.

Lewis, David. 1974. "Survival and identity". In The identities of persons. Berkeley and Los Angeles: University of California Press.

Misak, Cheryl. 1995. Verificationism: Its history and prospects. London and New York: Routledge.

Parfit, Derek. 1984. Reasons and persons. Oxford: Oxford University Press, Clarendon.

Perry, John, ed. 1975. Personal Identity. Berkeley: University of California Press.

Quine, Willard Van Orman. 1960. Word and object. Cambridge, Mass.: MIT Press.

Rorty, Amélie, ed. 1974. The identities of persons. Berkeley and Los Angeles: University of California Press.

Taylor, Charles. 1989. Sources of the Self. Cambridge, MA: Harvard University Press.

Wolf, Susan. 1986. "Self interest and interest in selves". Ethics 96:704-20.

 
BACK TO Ronnie de Sousa's Home Page