Patrice Maniglier (University of Essex) The Others Truths: Logic of Comparative Knowledge Departmental Seminar of the Department of Philosophy

of the University of Essex (17th December 2009)

Please be aware that this is only a draft which is not even fully referenced: it is not to be quoted as such, but rather to be discussed and improved... Comments, corrections, remarks and questions are more then welcome and can be addressed to me via email: pmanig@essex.ac.uk . I want to thank Fabian Freyehagen and David McNeil for their questions and suggestions in the course of many informal conversations we had on issues related to this paper. However, since they didnt read any written version of this paper, and certainly dont share the views presented here, they cant be held in any way responsible for all the weaknesses of the following argument.
And what I dont for the life of me make out, Chad pursued with resigned perplexity, is what you gain by it. Oh it would have taken his companion too long to say! Thats because you have, I verily believe, no imagination. Youve other qualities. But no imagination, dont you see? at all. I dare say. I do see. It was an idea in which Chad showed interest. But havent you yourself rather too much? Or rather ! So that after an instant, under this reproach and as if it were at last a fact really to escape from, Strether made his move for departure. (Henry James, The Ambassadors. London: Penguin Classics, 2003: p. 437)

Nothing illustrates more nicely the view that the very notion of truth is incompatible with the idea of its being relative, than this famous line by Pascal: Truth on this side of the Pyrenees, error on the other side. That this line is ironical, and means exactly the contrary of what is says, is even more apparent in its original context, which is actually plagiarized from Montaigne or, to speak more politely, implicitly referring to a passage from the Essays:
We see neither justice nor injustice which does not change its nature with change in climate. Three degrees of latitude reverse all jurisprudence; a meridian decides the truth. Fundamental laws change after a few years of possession; right has its epochs; the entry of Saturn into the Lion marks to

us the origin of such and such a crime. A strange justice that is bounded by a river! Truth on this side of the Pyrenees, error on the other side. Pascal, Penses, 294 (Edition Brunschvicg) 1 . Quelle bont est-ce que je voyois hyer en credit, et demain plus, et que le traict dune riviere faict crime ? Quelle verit que ces montaignes bornent, qui est mensonge au monde qui se tient au del ? (Montaigne, Essais, Livre II, Apologie de Raymond Sebond , GF, p. 245).

The idea is that one and the same proposition cannot be true only in certain circumstances, and become false in others, in other words, that truth cannot change. To say that truth is relative in the sense that what is true is certain circumstances becomes false in other seems simply to say that there is no truth, but only opinions. Of course, the fact that people disagree about what is true or false is certainly not enough to make it so that there is nothing which is true, but, if anything is to be true, it must be true in spite of the various opinions that people have about what is true. What is true now will always be true (if it is true indeed), and what is true here will be true everywhere (if it is true indeed). My aim in this paper is to interrogate this very assumption that relativity undermines the very idea of truth. But I am not so much interested in arguing in favour of what I will call the relativist thesis, which can be loosely paraphrased as holding that there is no ultimate criteria of what is right and wrong but only a plurality of assessments of the right and the wrong depending on conditions which cannot be themselves assessed from an overlooking impartial standpoint (external, as it were, to the arena of the dispute); I am more interested in arguing that the relativist postulate, far from dismissing the very idea of knowledge, is, on the contrary, an instrument for the production of a certain form of knowledge comparative anthropological knowledge. Of course, I will have to justify the use of the word knowledge in the previous sentence, and therefore to argue that it is possible to speak of truths based only on, or resulting solely from, the very assumption that there are no absolute and eternal truths and nothing beyond the variations of opinions or appearances. I suspect that you will be very keen on pointing immediately at the usual paradox against scepticism and relativism: that this is itself, by its own standard, only an opinion and cannot pretend to say the truth about opinions I will try to explain why I dont find too much difficulty in this and happily confess that relativism is indeed relative. But I will argue that relativism is not so much an epistemological thesis, as a methodological postulate, and I must confess immediately that I am more interested in the productivity of such a postulate in terms of its consequences than in its intrinsic absolute validity or in the question whether the consequences themselves can be
1

On ne voit presque rien de juste ou d'injuste, qui ne change de qualit, en changeant de climat. Trois degrs d'lvation du Ple renversent toute la Jurisprudence. Un Mridien dcide de la vrit, ou peu d'annes de possession. Les lois fondamentales changent. Le droit a ses poques. Plaisante justice qu'une rivire ou une Montaigne borne ! Vrit au de des Pyrnes, erreur au del.

proved to be absolutely valid. I will be happy if I find a standard of assessment of the right and the wrong which results uniquely from the assumption that there is nothing but a variation of assessments of the right and the wrong. But it must be said from the beginning that I will argue that relativism is not the idea that truths are relative to different contexts which would exist in themselves and be independent 2 , but rather the redefinition of everything that seems to be, and in particular of these contexts, in terms of how they relate to one another and can be seen as variants of one another. I would like to call this form of relativism affirmative or active relativism, and oppose it to what could be called negative or reactive relativism. Reactive relativism concludes that we must suspend our judgement about everything and hold to the general view that what is here is contingent in the sense that it could be entirely different (there is no necessary reason for it to be and to be the way it is, to paraphrase Leibniz), and cannot be grounded on any absolute necessity. Active relativism is not, or rather, not only, a negative statement about the impossibility of anything (i.e. the absolute foundation of any sort of value), but rather a positive method to generate new views on ourselves and things around us and to draw from the relativist postulate itself a criterion to redefine what there is as a variant of other possibilities of itself. The truth of what we are is to be found in the others truths.

1. Some arguments for the Relativistic Postulate. Even though I could simply consider relativism as a postulate and go on to analyse how it can operate as an instrument for a certain kind of knowledge, I thought it was at least useful, and probably also quite fun, to try and give some of the reasons why I think the relativist postulate is not such an unreasonable claim as it is often portrayed.

1.1. The Amazonian body and ours: In order to make the discussion immediately more concrete or dramatic, I will start with an example: in an article entitled The Gift and the Given, Three Nano-Essays on Kinship and Magic, the Brazilian anthropologist Eduardo Viveiros de Castro (to which the ideas presented here are deeply indebted, though I will introduce them in a very different

This definition of relativism is the only one which Chris Swoyer considers in his entry Relativism for the Stanford Encyclopedia of Philosophy (http://plato.stanford.edu/entries/relativism/, last accessed 16/12/09).

style), reports the following dialogue gathered on his fieldwork by his colleague Peter Gow among the Piro, an Amazonian people:
A mission schoolteacher in [the village of] Santa Clara was trying to convince a Piro woman to prepare food for her young child with boiled water. The woman replied, If we drink boiled water, we get diarrhoea. The schoolteacher scoffed, and said that the common infantile diarrhoea was caused by drinking unboiled water. Unmoved, the Piro woman replied, Perhaps for people from Lima this is true. But for us native people from here, boiled water gives us diarrhoea. Our bodies are different from your bodies. 3

This reply echoes another infamous one, given by a Zande man to the British anthropologist Evans-Pritchard, saying of Europeans: perhaps in their country people are not murdered by witches, but here they are. 4 Of course, these claims seem clearly wrong: the proposition the water should be boiled in order to prolong the life of the child (or at least to avoid diarrhoea) may be true in the world of the schoolteacher but not true in the world of this native Amazonian woman seems to be demonstrably false. The schoolteacher may have replied: Just try, Maam! And this would be a perfect illustration of the impossibility of holding that what is true here may not be true there. However, I suggest we should think of it in this way: what if it appeared that the kind of life that people have to live in order to prolong their life through boiling water rather than not was an altogether different life? What if, in order to accept the objective claim (which experience substantiates), this woman had not only to drop this particular belief, but also a huge set of beliefs, and, ultimately, to change the very sense it makes for her to live and to die? Wouldnt it make sense to say, in this case, that you havent presented to the Piro woman simply the truth about life in general, but rather a new definition of life, or, in the vocabulary I will favour, that you have not said anything true about the world, but asked her to renounce her world? Therefore, my first argument to support the claim of the Piro woman against that of the schoolteacher, that is, to support precisely her relativist claim but in terms understandable by us, is that the words life and death dont mean the same thing for her, so that to behave as scientific medicine says one should behave would imply not so much improving your life or prolonging it, as rather living a different kind of life. 5 (One way of
3

Peter Gow, Personal Communication to Eduardo Viveiros de Castro, quoted by Eduardo Viveiros de Castro, The Gift and the Given, Three Nano-Essays on Kinship and Magic, in Kinship and beyond: the Genealogical Model Reconsidered, Sandra Bamford and James Leach (eds.), Berghahan Books, 2009, p. 235. 4 EVANS-PRITCHARD, E., 1937, Witchcraft, Oracles and Magic among the Azande, Oxford : Oxford University Press, p. 540. 5 Now, I would probably have to go further and say that my present argument is precisely not what the Piro woman meant, the very idea of cultural relativity being itself a very cultural way of looking at the differences;

understanding what I am saying here is to relate it to what Heidegger argues about the ultimately hermeneutic constitution of the world: scientific knowledge itself is not just the discovery of what there is, but is itself grounded on a specific constitution of the world, a certain way of engaging with the world.) Now, you could accept all this, and nonetheless reply that it doesnt make the statement the water should be boiled in order to prolong the life of the child true in the world of the schoolteacher and false in the world of the Indian woman: it is just that the meanings of the words have changed. In other words, it is not the same truth, so that it makes no sense to say that the truth varies or that it changes. It is here that things become more complicated. I would argue that these two worlds do in fact intersect: they are talking about water, about child, about diarrhoea, etc. and they do understand one another to a certain extent. They can get along for a certain number of things and can interact relatively smoothly on many other occasions. In other words, there are times or rather circumstances when the semiotic identities in one world seem to be recognizable, without difficulty, as the semiotic identities of the other world. We dont need to say that they are talking of the same thing, in the sense of any independently observable reality; on the contrary, we are just saying that they make sense of what the other one makes sense of, and that there is no other way to judge the identities than how they construe them in their reciprocal relations: in a way, it is the fact itself a child is drinking water, which is equivocal, in the sense that it exists as a relevant possibility in both worlds, and this is the reason why they disagree. In a way, it is continuously and smoothly that the statement becomes false: we seem to be doing the same thing, bringing up children, avoiding sickness, drinking water, and then at one particular point it appears that we may be doing rather different things. The interesting point is that it is precisely thanks to (or because of) some disagreements of a particular kind that we come to realize that we in fact dont mean quite the same thing. But this disagreement is construed and measured in the very terms which come into doubt. This is the reason why I think we cannot so easily say that they just mean different

but at least my argument suffices to make the point about the necessity of understanding what she means. Eduardo Viveiros de Castro makes a much stronger argument, but nonetheless similar, when he remarks that the problem is not which theory about life and death, health and illness, that of the Piro woman or that of the schoolteacher, is true, but whether what the Piro woman propounds is indeed a theory on the same object, and even whether it is meant to be a theory at all, that is whether we understand what she means by body when she says: we have a different body: What the argument expresses is another objective fact: the fact that the Piro and Western concepts of body are different, not their respective biologies. (241). He means here that it is not two different theories on the same thing, but precisely two different ways of construing what it is that we have to make theories about, and even what it is to think in general and to be in relation to something to be thought about.

things: what they mean cannot be construed but in their own reciprocal terms. They are doomed to misunderstand their misunderstanding. But I may leave this question of the relation between truth and equivocation aside for the time being, since I will come back on it at length later, and just retain the idea that it is at least not absurd to say that, if, in order to accept that boiling water will prolong the life of your child, you may have to change the very definition you have of life, then you can rightfully say that it does not prolong the life of your child, but redefines it.

1.2. Are the wood sellers indeed selling wood? My second argument elaborates on the previous one at a more fundamental level, that is, at the level not only of what we think we have discovered to be true as a matter of fact, but rather of what we think cannot be disputed without the very possibility of establishing facts being jeopardized. I will again use an anecdote, but this time fictitious, since it is the nice little parable of the wood sellers imagined by Wittgenstein in the Remarks on the Foundations of Mathematics and commented here by John Callanan a few weeks ago. I remind you of the important passage:
149. Very well; but what if they piled the timber in heaps of arbitrary, varying height and then sold it at a price proportionate to the area covered by the piles? And what if they even justified this with the words: of course, if you buy more timber, you must pay more? 150. How could I show them thatas I should sayyou dont really buy more wood if you buy a pile covering a bigger area? I should, for instance, take a pile which was small by their ideas and, by laying the logs around, change it into a big one. This might convince them but perhaps they would say: Yes, now its a lot of wood. And costs more and that would be the end of the matter. We would presumably say in this case: they simply do not mean the same by a lot of wood and a little wood as we do; and they have quite a different system of payment from us.

What interests me in this passage is the fact that Wittgenstein shows us that we may come to realize that the very validity of our judgements about the right and the wrong is conditioned by the fact our not misidentifying the practice, that which those we judge are doing, in a way similar to our misidentification of what the Piro woman understands by body when she says that we and the Piros have different bodies. If, after having tried my little experiment to show them that they are wrong according to what I think to be their own principles, they still dont accept what seems to me evidence for their incoherence, I have two equally open possibilities: either I think they are just stupid and crazy, or I accept that more important than the problem whether they are right and wrong is the problem of what they

actually mean by a lot and a little. As we know, for Wittgenstein, this means how they actually use these words, that is, ultimately, what they actually do when they seem to do what I call selling woods, i.e. their practice. 6 I want to emphasize the fact that Wittgenstein does not say that the conduct of the wood sellers is simply incomprehensible and that this is the end of all matters, but, on the contrary, that it may become comprehensible if we were able to reconstruct the nature of their activity, their system of payment (and therefore that the very condition for anything to be comprehensible at all depends on the activity it is embedded in).7 In other words, I dont think that he simply wants to say that there may be people who think in a way completely different from our own, but rather that the question is misplaced and has to be rephrased: not whether the wood sellers are right or wrong, rational or irrational, but rather what it is that they are doing and whether we have ways to redefine our practices in such a way that we can engage into theirs and reciprocally. Wittgenstein remarks that up to this point we had supposed that they were actually doing the same thing as we ourselves usually do: paying. But here, it appears that this may have been an equivocation: what I took for a form of payment may be something quite different, while having some relation to what we call paying. This is exactly where the anthropological investigation begins. This remark, therefore, is not the end of all matters, but the beginning of another matter, the anthropological matter. Wittgenstein himself doesnt do much here to explain how we could reconstruct this alternative form of payment which he presumes, but he nonetheless firmly leads us to the point where anthropological questions arise, which is exactly the point where normative judgements fail. The remark about the fact that they may have quite a different system of payment from us is indeed typical of a good anthropological question: it suggests that this variation in the practice of counting may be understandable if we resituate it in the context of variations of other aspects of their practical life. It suggests that there is a point where we cannot severe our arithmetical tools from the practices in which we use them, and that, if we had a completely different system of payment, that is, ultimately, a different practice of justice, we may end up with a different sort of arithmetic, that is, different intuitions about the more and the less, the relations between our economic practices and our arithmetic practices becoming

For further occurrences of this concept of practice in Wittgenstein, see among many other ones, the paragraphs 198, 199 and 202 of the Logical Investigations. 7 Elaborating of what Callanan writes, we can remark that the presumption of comprehensibility, that is, of meaningfulness, is not grounded on some sort of general charity, but that it is quite the opposite: we are charitable just because they look like us, their activity resembles ours. We have no other reason to find what they do meaningful, but because it has some family resemblance with what we do.

more apparent if we could work about what it takes for the wood sellers to have different intuitions about the more and the less. 8 Consequently, what I take to be the important point here is not only the idea that normative questions are merely dependent on matters of fact which cannot themselves be assessed (Gosh! They are so different!), so that the only thing which could result from the confrontation with ununderstably alternative normative judgements is simply the conviction that our normative standards are ultimately contingent, but rather the idea that normative questions about the right and the wrong stop at one point and give way to anthropological questions which focus on the problem of the misidentification of practices and try to use what we have in common with alternative forms of life in order to understand better what should be modified in our practical life so that we understand better what at first seems ununderstandable. More important than understanding is the problem of even defining our own practice in relation to the other ones: it is quite striking that, even in this very brief passage, Wittgenstein does allude to the fact that we were not aware of the particular relations between our arithmetic theories and our economic system until we were confronted to a different understanding of a lot and a little: this is the critical part of cultural diversity: it helps us to become more aware of relations between different levels of things which seemed obvious for us. We discover first that evidences are not self-evident but rather co-evident; and second that the very identity of what we do can be reached through the untranslatability of our evidences In a way, Wittgenstein shows that there is no internal access to the identity of what we our doing and even to the question of the nature of our own activity: we cannot distance ourselves from the fact that we are counting, unless we discover that you can understand the very notion of count in a different ways, meaning by this nothing but the fact that something very similar to what we call counting can be in fact different given the practical context in which it is performed. In short, normative judgements depend on practical identities: I cannot judge somebody as doing wrong if I dont presume that we are engaged in the same activity. Therefore my judgements are always threatened by the possibility that what looks like what we usually do is in fact different from it, that is, by an equivocation similar to the one I suggested there was between the schoolteacher and the Piro woman. In other words, the real problem with human beings is not so much that they dont know what is ultimately right and wrong and have various opinions about that, but rather that, as God Himself said

This remark suggests that mathematics is a form of hermeneutics of our practical life. This hermeneutic definition of mathematics has been defended, among others, by Jean-Michel Salanskis, Hermneutique Formelle (Paris: CNRS Editions, 1998).

and these happen to be His last words: they know not what they do. Isnt that a good reason to forgive them? But do we know more what we are doing ourselves? Do we really understand what we mean by counting? Do we know how to characterize the identity of this activity, the way it relates to other aspects of our practical life? I doubt it, and the fact is that Wittgenstein agrees that this practical equivocation begins at home: this is the sense of many remarks in the Logical Investigations in which Wittgenstein argues that there is no way one can draw a rigid line between making errors in the application of a particular rule or following another rule. We find the same idea in this line from the Remarks on Colour: There is a continuum between an error in calculation and a different mode of calculation (293). And this is an immediate opportunity for me to say that anthropological problems are not only interesting because they suddenly confront us to the contingency of the way we do things here, but also because they may offer the only way we have to understand what it is that we are doing: the variations between the assessments of the right and the wrong dont simply challenge our own ones, but may also help us to characterize the identity of our practice. This is certainly the main message I would like to convey in this talk: we can do something else with cultural variety than try to overcome it in order to find universal standards of truths, or, on the contrary, mourn over it and lament about the contingency of our standards: we can use it to understand ourselves better, and to understand us only in relation to others, which we discover through truth-variations. In other words, if you could be forgiven until now for not knowing what you do, now that you have the opportunity to be confronted to your fellow ignorant human beings in normative clashes, you may have the opportunity to define what it is that you are doing by comparison or difference with the others. Now, of course, many questions arise. The first one is whether I can indeed find what in my practical form of life should be changed for this mode of calculation to make sense. This implies that I am capable of using my own practical categories to operate with and within them a variation which will modify them so that I will end up in another practical world. The question is here whether it is possible to enter into another practical world only using the resources of our given practical world. I will try to explain how this can be done later on. The second question is whether this remark by Wittgenstein supports any form of relativism at all, in the sense that it would undermine the idea that I am absolutely right to use the words a little and a lot in the way I normally do when I practice arithmetic. Indeed, one could argue that it would still be true that, if you want to do what we do here when we

sell and buy, you have to behave in this or that way, or mustnt do what the wood sellers do. It is true in all possible worlds, that, in our world, you have to judge this proposition true and this proposition wrong if this is what we call counting that you want to do. The fact that propositions are true only under certain conditions doesnt make them false in the absolute, but simply displaces the truth value onto the relations between the conditions and the conditioned. It is just like in logic: you can indeed change your axioms, and you will then have different theorems, but it is all the same true that this or that theorem necessarily follows from this or that set of axioms. It is still universally true that this is relatively true, and, therefore, truth itself remains absolute. In answer to this, we first have to recall what we just said, that is, that there is no guarantee that our practice is locked into itself in such a way that it cannot change and end up being the one of the wood sellers (and, hopefully, we may find under what conditions this may happen by trying to see what must be changed in our practical life for us to have these new intuitions). But even if this what not the case, even if we could argue that the right and the wrong are dependent on or relative to practical forms of life which cannot themselves change, so that it makes no sense to say that what is true may become untrue in a continuous way, it will still be very meaningless to say that what is right here is absolutely right. First it must be remarked that the comparison with the axiomatic mode of reasoning is odd here, since, if we are talking of logic, that is, of the theory of formally valid inference, we are talking of that which enables us to go from the premises to the consequences: if we accept that this can vary, as presumably Wittgenstein does, how can we argue that it is still true in all particular worlds that the relation between the contextual premises and the theoretical consequences are valid? True it may be in this particular world: if there is indeed no way my world can change and become a different world (contrarily to what Wittgenstein argues), then I can comfortably indulge in the local necessity I experiment here and not worry about its possibility of becoming false in any possible state of my world. But even then I have no reason to say it is absolutely true. Secondly, I think there is a sleight of hand going on here, which consists in forgetting what we have just said (that the concepts of right and wrong are only valid under the condition of the identity of the practices), as soon as we say it. For, if the words right and wrong only make sense in a particular practical context, how can we say immediately that the relation to the context will still be right in all possible contexts: what is this context from which we can judge the validity of the relation of a normative claim to its context?

Wouldnt we need here something like a context of all contexts (a world of all the worlds)? But how do we access the context of all contexts, if this ever makes sense? This leads me to my third remark: if we dont have an a priori given global context, the only way we can check that a contextual inference is true in another context, relies entirely on our capacity to prove that we can pass from one context to the other, while preserving the identity of what is being assessed, that is, showing what must be varied when we pass across this threshold which separates the worlds: then indeed, it would be true to say that the fact that it is true in world A would be itself true in world B, because we know precisely how to reconstruct world A starting from world B, in the sense that we know the practical conditions which must be modified for us to pass from one to the other, and thus can locate unequivocally the proposition in this reconstructed world. In other words, nothing is globally true unless we have the rules of passage or transformation between the contexts: the global context is nothing but the set of these rules of conversion of one context into the other one. If I dont have such rules, it makes no sense to say that the fact that it is true here is itself true everywhere, quite simply because there is no stable relation between these worlds. Again, this doesnt mean that I have to abandon my rules: it just explicates the conditions of their validity. Lets then accept at least for the sake of the argument that we have good reasons to posit the relativist postulate and say that our assessments of the right and the wrong are dependent on practical worlds which themselves cannot be assessed according to any normative standards. Behind our standards of truth, we then find problems of meaning; and behind problems of meaning, problem of identification of the practices. Now the question is: can we use this variety of normative standards in order to understand better the identity of the practices which underlie them? But before getting into this, I must probably answer to a general worry, which is that the very postulation of the relativist postulate is itself selfrefuting.

1.3. The self-refutation objection: whose foot am I supposed to shoot in? The idea that radical relativism is self-refuting is a very old one. It can be found in Platos discussion of Protagoras claim that man is the measure of all things, in Thaetetus, (171a-b).
Socrates: Well then, shall we say that in such a case, you opinion is true to you but false to the myriads?

Theodorus: That seems to be the inevitable deducation. Socrates: And what of Protagoras himself? If neither he himself thought, nor people in general think, as indeed they do not, that man is the measure of all things, is it not inevitable that the truth which he wrote is true to no one? But if he himself thought it was true, and people in genera do not agree with him, in the first pace you know that it is just so much more false than true as the number of those who do not believe it is greater than the number of those who do. Theodorus: Necessarily, if it is to be true or false according to each individual opinion. Socrates: Secondly, it involves this, which is a very pretty result; he concedes about his own opinion the truth of the opinion of those who disagree with him and think that his opinion is false, since he grants that the opinions of all men are true. Theodorus: Certainly. Socrates: Then would he not be conceding that his own opinion is false, if he grants that the opinion of those who think he is in error is true? Theodorus: Necessarily. 9

Well, granted: relativism would be self-refuting indeed, if it said that it is absolutely true that truth is relative to practical contexts, that is, that it is true in all possible worlds, or, rather, independently of any world and of any way of actually construing the passage from one world to the other. For instance, that truth is relative for the Piro woman or the woodsellers just in the same sense as it is relative for me, without taking into consideration precisely the relations between these worlds, as from an overlooking standpoint. But we dont need to say that truth-value relativism is absolutely true: we can simply say that it is relatively true, that is: given the context in which we live and think (given the sort of experiences that I put forward above), it is more plausible to think that every truth is relative, by which we ultimately mean that the identity of each truth is defined by the way it could become false (the semantic contents of each truth are relative to one another). We dont even need to say that there are contexts (or, as Chris Swoyer puts it, frameworks) to which truths are relative, as if these were independent variables which we must simply discover as they are, waiting for us on the other side of the curtain which separates us from the truth. We can simply say that such frameworks conditioning what we hold to be true appear in certain circumstances, more specifically in circumstances of cultural encounter or historical change when it seems that what seemed to be true become seemingly false, and that we have no access to those conditions but through their reciprocal relations as we can reconstruct them on the basis of our misunderstandings.

Plato, Theaetetus Sophist, trans. Loeb. Harvard/London: Harvard University Press, 1987: p.109-111. This argument is abundantly supported in the entry on Relativism written by Chris Swoyer for the Stanford Encyclopedia of Philosophy (http://plato.stanford.edu/entries/relativism/, last accessed 16/12/09).

The question is whether I am entitled to use the word truth to designate something which is only relatively true. If by truth you mean something which is absolutely valid and cannot be relativized out of principle, then, indeed, there is nothing like a relative truth. But this is just a very poor tautology and you have put the answer in the question. Then indeed, it is simply of question of word, and I can use another word if you want to reserve the notion of truth for your cherished chimera: I can call mine item of scientific knowledge if you prefer. But if you want the question to remain open, as to whether the truth can be said in an interesting way relative or absolute, we have to ask ourselves whether what we expect from the concept of truth can be obtained equally (and maybe even better) by construing it as relative or as absolute. In other words, truth must be defined in continuity with what we already do when we pretend to know things: knowledge is not defined by the notion of truth, but it is quite the opposite: the notion of truth is defined by the standard practices of knowledge. For this reason, comparative knowledge (anthropology) must look like (at least in some respects) other scientific practices. To cut very short what could be a very long argument, and speaking very loosely, I would say that we expect from a truth procedure two things: that it enables us to create new opinions (it changes the way we look at things, that is, the world in which we live), and these new opinions appear to be more compelling than the previous ones. In other words, we only need to preserve some sense of refutation on a local basis, meaning that, given the evidence that we have here (how things look like from here), some new elements can be created, which also become more apparent (more compelling) than other ones. Therefore, if the statement truth is relative, understood as it seems that what seems to be true can become untrue in certain circumstances, is itself said to be relative, it means that the statement what seems to be more compelling may appear to be less compelling in certain circumstances is itself more compelling than its opposite in the circumstance in which we find ourselves, while we dont rule out a priori the possibility that there may be other circumstances in which it will be found less compelling. I dont see what is self-contradictory in this idea: we plainly agree that truth may appear to be absolute in certain contexts, while, in the context in which we are, it appears to be relative, maybe for reasons which can become clear if we understand better what should be changed in the way we think in order for us to believe that truth is absolute... The real question is rather whether this very assumption, the relativist assumption, is itself an endpoint in the search for truths, or if it provides methodical procedures to create new truths. If it does, then relativism, far from being incompatible with the idea of knowledge, opens up to a certain kind of knowledge.

To the objection that, if the relativists claim that truth is relative is only true relative to his framework and can be false in other, perhaps equally good, frameworks, then you have no reason to even care about the relativists (perhaps rather idiosyncratic or parochial) framework, as Chris Swoyer argues in his article from the Stanford Encyclopedia of Philosophy, I have just one simple answer: because it may be yours too The question, in front of any philosophical or even theoretical claim, is whether it operates strongly enough on your way of thinking to tear you away from where you were and bring you to somewhere else though as if from some sort of internal or immanent necessity. I must say I have some hard times understanding what else is being asked here: are we supposed to care only about those claims that would be true even if we had no access to the reasons why they may seem true? Isnt it more reasonable to simply say that we have indeed nothing but appearances and apparent standards for ranking these appearances, so that we will simply try to go for the most compelling ones? You may object that, we shouldnt then bother about doubting at all in the first place, since we want to go for the strongest argument anyway: but the problem is that we dont have the choice to doubt or not to doubt, we are confronted to situations which make us think that opinions are doubtful, situations of the kind I referred to above. Of course, if you dont share this context, or just dont understand these arguments, you have absolutely no reason to doubt about them, and I would even plainly agree that you are right in not doing so: why would you do it indeed? I dont expect you to believe me just because you like me. To this sort of objections to the relativist view, I am tempted to reply as Descartes did when Father Mersenne, in the Second Objections, objected that indubitability may not be enough to prove that something is true, precisely because it made truth dependent on the way we think, by defining it as that which seems impossible to deny. He replied that this definition of truth provided everything that we could reasonably want.
What is it to us that someone may make out that the perception whose truth we are so firmly convinced of may appear false to God or an angel, so that it is, absolutely speaking, false? Why should this alleged absolute falsity bother us, since we neither believe in it nor have even the smallest suspicion of it? (Replies 2, AT 7:144-45)

I am no more relativist than Descartes when I say that the only thing which seem certain to me is that my certitudes may change; I may even be more coherent than this distinguished author and also more aware, because of the evolutions of the framework he himself contributed to set up, that there is no single way to define this capacity of perceiving immanent necessities which he called reason. But like him, I am speaking for myself

indeed, and all I can do is to be as true as possible as to how things seem to me 10 , and I will be happy if I find some appearances (opinions, beliefs) which are more compelling than other ones nonetheless: if this appearance is that appearances can change, I am not self-refuting myself, but self confirming myself. When confronted to different views which seem to oppose mine, the best I can do is to see whether I can change my framework so that what seems evident for them may not also become evident for me, and grant that, first, it is possible that this conversion may happen, and that, second, it is even possible that from another standpoint I come to believe that our beliefs are not different, in other words that it doesnt seem that what seems to you and what seems to me seem different. It seems to me that, given the kind of examples that I have provided above, it should appear to everyone, making the same assumptions as I am making and confronted to the same evidence, that this seems the right conclusion to draw, but I happily confess that I will not try to convince anyone on any other basis than how things seem to him or her, even if it is in the end to make him or her change their mind, just as I changed mine I must confess that I think that ones own foot is the only worth shooting in. This whole discussion reminds me of the problem of scepticism, and whether scepticism is or not self-refuting. The most violent affirmation of the self-refuting nature of radical sceptical claims is to be found in Spinoza, in the Tractatus de Intellecti Emendatione.
If they affirm or doubt something, they do not know that they are doubting or affirming. They say that they know nothing, and they say that they are ignorant of this very fact of knowing nothing. And they do not even say this without qualification; for they are afraid that, in saying they know nothing, they are declaring that they exist, so that in the end they have to maintain silence lest they should perchance say something that has the savour of truth. () So they must be regarded as automata, completely lacking in mind. (Treatise on the Emendation of the Intellect, 47-48, trans. S. Shirley (Indianapolis: Hackett Publishing, 1992: p. 243-244)

I must say I disagree with the idea that radical doubt is itself impossible. For, I can very easily say that I do have reasons to doubt, but that I also doubt of their validity, meaning that I doubt whether I am right to doubt or not. In other words: I dont know whether I am right in not finding anything absolutely right or wrong. What is wrong in that? Of course, if I were to say that I doubt the very fact that I am doubting now, this would be more problematic.

It may be worth mentioning that this is also the same principle that Socrates invokes against Protagoras: Most likely, though, he, being older, is wiser than we, and if, for example, he should emerge from the ground, here at our fee, if only as far as the neck, he would prove abundantly that I was making a fool of myself by my talk, in all probability, and you by agreeing with me; then he would sink down and be off at a run. But we, I suppose, must depend on ourselves, such as we are, and must say just what we think. (171D, p. 113).

10

But I am simply saying that I consider my reasons for doubting themselves doubtful. If now you want to do as Descartes, and shout that this confirms at least that, as a doubting activity, you exist, I have no problem with this, as long as you dont separate yourself from the reasons you have of doubting, that is of believing that you may become different from what you seem to yourself, and that you ultimately realize that this commits you to accept that the verb to exist doesnt mean anything else than to be able to become different I hope this is enough to support the idea that the relativist postulate is not immediately self-refuting. However, as I said in the introduction, I am not so much interested in the question whether it is true that truth is relative or not, as in the question: what can we learn from the postulation of this hypothesis? Is it possible to build up a form of knowledge on the basis of the relativist postulate? (And again: by knowledge I dont understand more than a symbolic mechanism which has the capacity of producing new appearances which seem more compelling according to contextual falsifiability criteria.) I dont pretend that there is no other way of trying to produce some knowledge about the world (something which is more worth holding than other things), but only that it is possible to construct a kind of knowledge on the very idea of relativity. In other words, I would like to take the relativist claim as a postulate. My attitude is more similar to Descartes, assuming the most extreme doubt in order to turn it against itself and base on it its first truth and then the entire system of knowledge, than to a negative form of relativism: I want to use the most extreme doubt, which is the idea that not only we cannot be sure that what seems true to us is indeed true or not, but also that what seems to be true to us may happen to seem to be untrue, to see whether it is not possible to build on its very basis another kind of knowledge. It actually seems to be a rather consistent feature of Modernity (very apparent in mathematics) that, contrarily to the image we often have of knowledge from outside, it advances by making less and less assumptions, by increasing the number of things we can be called into doubt, by depriving itself of more and more means and security, and trying to show that, even in these extreme conditions, we are still capable of finding some things more worth thinking than others. I think the modern subject of knowledge is very similar to Becketts characters: we have no leg, we have no arms, we have no eyes, no ears, no mouth, we dont even whether we are anywhere and maybe not even whether we are or not, and yet, with so little, we can still move We shouldnt measure the quality of a thought by how good it is at not shooting in ones own foot, but rather by how many bullets you can shoot in your own foot and still stand up and even run faster.

2. Consequences of the Relativistic Postulate: Epistemology of anthropological knowledge: For now on, I will then dismiss all the absolutist claims and stick to my relativist postulate: I will accept nothing as true than the idea that what seems true here may become false in circumstances which I dont even yet understand, or, in other words, that there is a plurality of assessments about right and wrong which cannot themselves be assessed, so much so that these different opinions are all that there is. And, rather than trying to escape from the changing, movable quicksand of opinions, in search of something which would be altogether different from an opinion, I will try to see whether it is not possible to build a kind of knowledge on this diversity itself, emerging as it were from its own absence of foundation. The advantage of such a decision is that it is minimal: it is nothing but what is actually given to everyone I can think to be in relation with: there are different ways of assessing what is right and what is wrong; any attempt at ranking them is itself an opinion and, practically, an opinion is nothing but such a ranking. Now the question is whether this refusal of any hierarchy, this position of a radical equality between the opinions, may not be itself enough to produce, as it were from its own demands, a way of creating new opinions, new appearances which will then appear to be more strongly motivated than others from the only perspective of the relativist postulate.

2.1. Of the things of which we may doubt that we are even in doubt about: It is likely that many will be tempted to refuse what I just said, that is, that the relativity or, rather, the mutability of opinions, is indeed a given, and will accuse me of introducing a strong supplementary claim when I claim that that the only thing I hold true is that what is being held true can change. After all, it doesnt seem to everybody that all the opinions are equal or that they could end up holding true what they held false. I would like to argue first that my postulate is minimal, and that it is nothing but what everybody must accept in order even to give sense to the very idea of truth. My postulate simply says that the way we assess the right and the wrong can themselves change. My argument for saying this is two-pronged: first I think that there is no reason to care about the truth but because we are in doubt; second, I think that we have no reason to be in doubt but because we have experienced that what we held true (our certitudes, our evidences) changed

or that we were confronted to different and non-assessable assessments of the right and the wrong. Even Descartes didnt doubt just because he was a particularly critical mind, but because he was confronted to a diversity of opinions in the guise of the conflict of authorities between his masters at school and between the different customs through which he travelled:
Considering how many opinions there can be about the very same matter that are held by learned people without there being the possibility of more than one opinion being true, I deemed everything that was merely probable to be well nigh false. () It is true that, so long as I merely considered the customs of other men, I found hardly anything there about which to be confident, and that I noticed there was about as much diversity as I had previously found among the opinions of philosophers. Thus the greatest profit I derived from this was that, on seeing many things that, although they seem to us very extravagant and ridiculous, do not cease to be commonly accepted and approved among other great peoples, I learned not to believe anything too firmly of which I had been persuaded only by example and custom. (Discourse on Method, Part One, AT, 9-10; op. cit., p. 50)

Reading and travelling, this is what put Descartes in doubt. Spinoza makes a similar point when he remarks that a man who would have been confronted in his life only to one image, the image of a winged horse, would then have no reason to doubt of its existence. In sort, what makes our opinions doubtful is nothing but the experience we have of their changeability, that is, of the possibility of thinking differently.
I deny that we have free power to suspend judgment. For when we say that someone suspends judgment, we are saying only that he sees that he is not adequately perceiving the thing. So suspension of judgment is really a perception, not free will. To understand this more clearly, let us conceive a boy imagining a winged horse and having no other perception. Since this imagining involves the existence of a horse (Cor. Pr. 17, II), and the boy perceives nothing to annul the existence of the horse, he will necessarily regard the horse as present and he will not be able to doubt its existence, although he is not certain of it. (II, 49, Scholium, p. 99)

However, while Descartes concluded from this the necessity of retreating within himself and taking sides only for what seems indubitable to him, I would like to stick to this simple given, the variability of opinions itself (or the awareness of the possibility that what seems indubitable here and now may appear dubitable), and, instead of trying to overcome it in order to find something which will be, ultimately, nothing but just another opinion (as unshakable as it may be according to Descartes), I would like to see if I cannot find in it some guidance for what should be preferred. I therefore firmly stick to this only belief: the only thing I hold true is that it is possible to hold false what I hold true. I am like those wanderers that Descartes pictures in the Discourse on Method (Part III), lost in the forest and wondering how they could get out of darkness, or this man, presented by Kant in How to orient ourselves in thinking, who has to go back to his house in a city where all the lights are switched off. I

have nothing but the different claims made here and there or spurting from myself about the path I should take. Many voices arise claiming: You could believe this, or you could believe that, but I stick to my only reliable belief for now which is that there is nothing decisively reliable. The question is: am I condemned either to choose one of these at random and stick to it as Descartes suggested, or to die here in the middle of the dark night and in the iridescent forest of opinions? Some very good intentioned voices suggest that, given my personal obsession with this relativist postulate, I may want to try to accept in this concert of opinions only the parts which are recurrent in all of them: I will hold nothing true but what is identical in all the beliefs, since I would then have no reason to call them in doubt. There are two reasons why this suggestion is unacceptable. The first one is that my postulate is precisely that any truth may be changed into an error: as long as I have no better reasons to abandon it, I cannot avoid thinking that each claim about the right and the wrong is contingent, and particularly in this case since the mere fact that I dont hear discordant voices doesnt mean that there may not exist in other parts of the forest or of the city. I know that this happened to me in the past, and now I suspect that all these voices may be as many evil geniuses scattered all around to deceive me. Havent I heard that there have been also many voices (i.e. cultures) in the past of which I know nothing since they have disappeared without leaving any trace: how do I know whether they didnt have a different view on the matters I ponder? But there is a more fundamental reason. Even if I was sure that I had made an exhaustive census of all the possible voices, it would still be possible that this lowest common denominator is itself subjected to various interpretations. Lets imagine for instance that all the voices seem to share something, which is the idea that certain sexual relations within the family must be forbidden. Of course, they differ dramatically as to which family relations must be banned (some tell me that it is my duty to marry my cousin, some on the contrary that this is exactly what must be avoided), but they all have some interdiction. However, I mustnt be too quick to rejoice and believe that I have hit a universal, which in this case would be called the prohibition of incest. Not only because it wouldnt be properly speaking a universal (as we know from Hume), but also because I have to be sure first that what all these voices call sexual, family and ban is identical. A little attention will show me that this is not the case. Here is a voice, for instance, which I will call the voice from the Trobriand (filtered through Bronislaw Manilowski), who says that the father is not a member of the family, but rather a foreigner who doesnt even take part in the conception of the children:

as a matter of fact, the word being used to describe a sexual relation he may have with a daughter is the same word as the one they use to describe a relation he would have with somebody else than his wife, that is a word which should be translated as adultery rather than incest. Similarly, I learn that some other voices understand by punishment not the fact that some centralized authority imposes some harm to the offender, but rather the fact that they immolate some other people which, according to them, are part of the general problem revealed by the crime. As for sexuality, another voice, Michel Foucault, tells me that it is a very recent concept and that even Rousseau wouldnt have understood what I meant by that. This is a point actually very nicely made in various contributions by the late American anthropologist Clifford Geertz, from which I will borrow this example which shows the emptiness of such lowest common denominators:
Zuni culture prizes restraint, while Kwakiutl culture encourages exhibitionism on the part of the individual. These are contrasting values, but in adhering to them the Zuni and Kwakiutl show their allegiance to a universal value; the prizing of the distinctive norms of ones culture. (C. Kluckhon, Culture and Behaviour, New York, 1962, p. 280, quoted in Clifford Geertz, The interpretation of cultures, Hutchinson and Co, London, 1975, p. 41).

The problem with such claims is that the very meaning of this prizing in general cannot be separated from its particular content: it is pointless to characterize something as being one moral system in general for what it means to be moral is precisely part of what our particular moral system tells us, or, to put it more concisely, what it is to relate to moral values also depends on the moral values we relate to. Same for less obvious cases: to say that all human beings use shelters or that they all have a certain relation to the dead, or that they all have the capacity of speaking, is perfectly useless, not simply because they build different sorts of shelter, mourn in different ways or speak different languages, but because the very characterization of these as shelters, mourning or language, is precisely the point in contention: it is impossible to separate the common characteristic from its particular forms. We understand by mourning nothing other than: to do something similar to what we do when we mourn: isnt the equivocation immediately obvious? Similarly, our very definition of what it is to speak is not unrelated to the very way we use our language, etc. Even to say that every human group has a culture or that all are differentiated from one another by their culture is perfectly equivocal and is a way of projecting onto other forms of human life our understanding of the way we relate to them, for which we use the concept culture, which is in fact not adequate to the way they relate to themselves and others. 11 The truth is that human
11

See Roy Wagner, The Invention of Culture, The University of Chicago Press, 1981.

beings differ precisely in the way they identify what they have in common! The common is the point of division and misunderstanding. Therefore the more abstract I will be, the more equivocal I will be. This remark makes me realize that I dont even possess the good measure of the differences and similarities between the various opinions I am considering, so that the difference between the opinions is itself a matter of opinion. For instance, if I remark that the claim no individual male must wear a skirt becomes untrue, say, in Scotland, I am already using concepts that may not well characterize what I am talking about: is a kilt a skirt? To take a more erudite example: if I want to compare the different forms of marriage existing around the world, I have to presuppose that, even though the rules determining who can or must marry who do vary considerably, the definition of marriage is stable; otherwise, my map of the differences wouldnt make sense. But, as Edmund Leach convincingly argued, the problem is that what is being understood and done under the heading of marriage around the world, while it may look like what we do, designates social practices which dont have much in common: here it only concerns the establishment of sexual rights, there it has absolutely no sexual dimension but simply refers to property rights, and in yet another place it is not marked by any sort of ceremony but very similar to what we call cohabitation, etc. 12 It then appears that it is not only the forms of marriage which vary but also the very definition of what varies. Of course, this sort of remarks abounds in the anthropological literature, and I will come back later on a more in-depth example. Suffices here to conclude that I have to darken again my already dark night and grant that, if I really want to stick to the given, I have to accept that the diversity of opinions is itself a diversity of the perception of the differences and identities between opinions: is given a diversity a way of construing this very diversity itself.

2.2. Of the Other and that it is better known than myself: the comparative intuition. It now seems that I am utterly lost and that nothing will ever be able to take me out of my deep night. However, the supplementary twist I have just given to the difficulty provides me in fact with the insight of a solution. Instead of trying to find out claims which would be universally valid, i.e. valid in all possible worlds, which seems impossible given the intrinsically equivocal nature of every concept, why dont I try to recontexualize each claim

12

Edmund Leach, Rethinking Anthropology (London : Athlone, 1971).

which pretend to universality, and more particularly those claims which pretend to give a measure of the similarities and differences between opinions, i.e. the comparative concepts, in order to show how they depend on and express the very differences that they obliterate or conflate? Intuitively, this means that I may try to localize the categorical grid that I myself use within the set of differences which it tends to obliterate. I thus propose to myself this new task: I will try to contextualize the very comparative grid I use by realizing the equivocations (the ambiguities, the misunderstandings) and having the meanings (or practices as a matter of fact) progressively diverge from one another, thus attempting at characterizing more precisely the differences by the way this seemingly common feature has in fact different senses when it is taken in a field of other differences which are not immediately apparent. For instance, instead of trying to use the concept of marriage (or, as a matter of fact, of truth, and even of opinion, worldview, culture, etc.) in order to compare different forms of marriages, I will try to redefine the very idea marriage by the relation between, on the one hand, the variations in the concept of marriage which I can reconstruct on the basis of the variations of opinions, and, on the other hand, other variations which situate, localize, or contextualize my concept of marriage. This means that, from now on, I will not accept anything unless it has been redefined as a variant. Since I have no other certainty than certainties change, even in the characterization of their very identities, I will now try to see whether it is possible to redefine every certainty as a variant. The idea that truth is relative here is not that it depends on something else, which one may call culture, practice, etc., which would be itself invariant, but rather that this something else is defined by the differential relation with the other truths. What do I call a variant? A variant is an entity whose identity is entirely defined by the way it could be different, which means that its identity is reducible to its position in a group of transformations, that is, in a field of objects which are related to one another by differential features only, and thus can be said to be alternative possibilities of one another. A variant is not a variable: a variable is not a term defined by its relation of transformation with other terms: it is the graphic substitute for a set of values which are in a determinate relation to another set of values. For instance, the probability of lung cancer is a variable which depends (among other things) on this other variable which is whether you smoke or not. The probability of lung cancer is not here defined by its differential relation to other objects, say, the other forms of cancers, no more than each value of this probability is defined by its differential relation to all the other ones in a field of transformation. A phoneme, on the

contrary, is a variant because it is only defined by the way it can be substituted by other phonemes. We usually use the term variant in a rather weak way: it consists in imagining that we are given a set of objects characterized by a certain number of properties, and that we call variants of the same type all the objects which share a certain number of properties but differ on the basis of other properties. Thus, for example, a text will be said to have variants in the sense that different texts are identical in most of their parts, but some sentences or passages differ. But this definition of the variant is unpracticable for me since it would require that I accept the identity of the types, and I have said that the types are precisely varying through the equivocations. I thus must define a variant as that which is entirely defined by its differential relations to other terms, and conversely, and must not accept any identity which is not the position in a group of transformations. Far from defining a variant in relation to a type, I will on the contrary redefine what I think of as types as determinate variants. I therefore posit the first rule of my method, which is nothing but the comparative method: never to accept any identity which cannot be redefined as a variant. Or, if I want to reinforce the relation between this and the problem of truth with which I started: only accept as true that which can be redefined as a determinate variant of that which could also be accepted as true. We can also give an ontological form to my first rule, meaning by this that I commit myself to a particular ontology: never to accept as existent anything which cannot be redefined as a variant of other possibilities of itself. To put it more dramatically: the only truth we can reach about what there is for us now (our situation, our world) is given by what there could be instead; or, more precisely, the only truth of what we are is given by what we can become (by how different we could be): the truth of what is actual is given by the alternative possibilities of this particular world which is actual. The possible, or rather, the virtual, is the real. Nothing truly exists but possibilities codetermined by the way they alternate to one another. Let me summarize what I just said: I held fast to the sole idea that there are different assessments of the right and the wrong. But this postulate brought me to accept that there are also different ways of perceiving the similarities and differences between these assessments. I then decided that I will refuse all global identity and will try instead to redefine every identity in terms of the differences it in fact mixed up (obliterated); I will then treat all types as equivocal terms and redefine them as variants in relation to the alternative interpretations of themselves which as types they obliterate. This process obviously implies that I relativize my own categorical grid (my way of making identities and differences) in the sense that I redefine

it by its position in the map of alternative ways of making identities and differences which it itself construes in the process of its self-relativization. The comparing instance is not itself outside of the field of what it compares, but the real upshot of comparison is precisely to situate my kind of knowledge in other variants of itself. Comparative knowledge is situated knowledge, but situated knowledge is a knowledge which redefines itself by its relation to alternative forms of itself. To compare means: to try to experience how the differences compared are themselves differently comparing and to try to redefine everything, and firstly oneself, as a variant of one another. This, from an anthropological point of view, means that what I have to compare are different systems of comparison. As Lvi-Strauss nicely put it, anthropology is the social science by the observed 13 . It is the knowledge of the others knowledge. This clearly implies that it is not simply an objective knowledge having to produce accurate descriptions of what is in front of us, but rather a critical knowledge which has to resituate the very subject of knowledge in the field of what it knows. Now, you will maybe impatiently say: how is that possible, since you just said that you didnt want to exclude the possibility that, whatever identity or difference you make, it can itself vary or be made differently? But there are two different questions here. The first question is whether it is possible indeed to experience a variation of ones own comparative grid on the basis of the differences which it makes locally perceptible, that is, whether it is possible to go from one regime of identities and differences to another alternative one, and therefore to the possibility of redefining each one by the variations which are necessary for going from the former to the latter, in a controlled way (that using a locally efficient refutability procedure). Here the question is whether it is indeed possible, simply thinkable, to use the only tool which is given to us, i.e. our comparative grid, to reach as it were continuously another, alternative, one, and therefore to relativize ourselves within the field of their differences. To say it again differently: the question is whether it is possible to construct, within my own thought, or my own language, an alternative system of thought, or an alternative language, from which I could look at what I used to be as it were from the standpoint of what I could become, so that I can redefine what I was as a variant of what I have become. I will try to show that there is no impossibility in this and that it is perfectly possible to use a system of differences and identities in order to experience within it a

13

Claude, Lvi-Strauss, The Scope of Anthropology.

variation of itself which is itself in a methodical way (constrained by a locally operational refutability procedure). Another question is whether I can be sure that the differences I use to do this, and the differences I end up with are indeed the right ones. But I dont have to answer to this question since my postulate is precisely that this question doesnt make sense: this is again a way of asking for absolute identities and absolute differences. I dont care too much whether the map of variants I will end up with is indeed the unique right one in the absolute sense; my question is rather whether the relativist postulate left me absolutely bereft of any criterion, so that, if you are a relativist, then just anything goes. If I can prove that the process sketched above is indeed possible, then it seems that not just anything goes: I have a criterion, never accept anything as true what hasnt been redefined as a variant, and it will certainly change the way I look and think one must look at the given (i.e. the variety of the ways of looking at one another). I can build on the relativist postulate a certain kind of knowledge, and whether this knowledge is absolutely true or not is not as important as the new descriptions which will be produced by trying to stick to the first rule of the comparative method.

2.3. Of anthropology, that it exists and can be illustrated through the example of kinship: To show how it is possible that an extension of my own categorical grid can result in an alternative one, I will analyse very sketchily an example taken from the anthropological literature, and more precisely from this founding moment in modern anthropology which is the work started by the American anthropologist Henry Lewis Morgan on kinship terminologies in the mid nineteenth century and systematized through the British and French schools of 20th Century anthropology. However, to be true to my little fable of the traveller lost in the dark forest of the night, I will still proceed in my Cartesian style. It makes no doubt that I must start with the differences which appear to me, within what I hope to be able to redefine ultimately as a variant, that is, my categorical grid. If I want to recharacterize (i.e. end up characterizing differently) the difference between my categorical grid and alternative ones (my voice and other voices), I cannot do it otherwise than by using my categorical grid. As Roy Wagner puts it: every understanding of another culture is an experiment with our own, The Invention of Culture (p. 12), or, as he also says, an extension (ibidem). I am thus given with resemblances and differences between the claims made by the different voices as they appear to me. For instance, I recognize a family

resemblance between those differences which characterize apparently for me and for other voices too what I call, precisely, a family. I mean by this that I perceive a resemblance in the ways by which these voices differentiate their behaviours in relation to the members of their family and mine. I also perceive however that one of these voices, the Iroquois voice, organizes these (differentiated) relations in a way quite different from mine: for instance, it says that it is not the name of the father that must be inherited by the children, but rather the name of the mother. I thus believe that the relevant differential feature between them and me is the ascription of the rule of descent or filiation, either on the side of the mother or on the side of the father. And since I happen to believe that this very idea of the name of the father is nothing but the remnant of a patriarchal society, I am quite happy to say that patriarchy is perfectly contingent and that there are voices which do not hold to this accursed view. Alas, I quickly come to realize that I have misperceived the similarities and dissimilarities (by which I just mean that the characterization of the differences between appearances I have just given doesnt do justice to the variations I can indeed perceive). Indeed, by listening to this voice more carefully or more at length, I understand that the word mother doesnt only apply to the woman who gave birth to the child, but also to her sister, the maternal aunt, and that grand-mother similarly designate the sister of the woman I would call the grand-mother, so that the daughter of this woman is in fact a mother too, and so on up in the genealogical tree, so much so that the Iroquois voice claim that each individual, instead of having one mother, has a lot of mothers. In short, I realize that there is here an equivocation. Reflecting on how I could make sense of this equivocation in the field of my appearances, that is what I would need to change in (my system of) beliefs so that it would appear to me to that each individual has many mothers, etc., it seems to me that it has to do with the fact that this voice doesnt in fact differentiate the family by the nuclear cell of the parents and the children as I do: this cell is rather irrelevant to the constitution of the kin, since the voice defines it by the group of individuals related to one another only by parents of the same sex (for which I have no ready made terms, which means that I can use their own, as some have done with mana or taboo, or find a word convey a meaning which seems very similar in a voice closer to me, in this case the Latin voice: clan, or lineage). While I thought the Iroquois voice was speaking of families, it was in truth speaking of clans. It thus appears to me now that the best way to characterize the difference between them and me has not so much to do with whether the rule of descent passes through the mother or the father, as, rather, with whether the relation between the direct line (for instance father-son) and the

lateral line (uncle-nephew) is treated differently (and then the uncle will not be a father) or identically (and the mother will be an aunt). This introduces within my world a variation which I had absolutely no idea was possible, and forces me to envisage a possibility which, without being in principle unthinkable, was necessarily unthought given the nature of my way of making differences and identities. Of course, I still express this differential parameter in my own terms (direct line, lateral line), which are potentially equivocal, but it doesnt change the fact that I have created a new differentiating feature out of my own categories, and that I never thought I could characterize myself by such a difference. This difference didnt appear in my world, and I now have to redefine myself as being characteristic because of this differential property which had no sense for me. Now, endowed with this new principle of differentiation and comparison, I will see whether I can characterize my difference with other voices and the difference between one another in the same terms, and whether it is possible to redefine each of the worlds which equivocally appear in mine (as well as mine) as variants along the line of this differential feature which maybe none of them recognized from inside as a relevant feature of their identity. This is what is know as the theory of descent group particularly favoured by British anthropology in the first half of the 20th century, establishing five types of kinship systems: unilineal systems for which the belonging to one kin group depends either on one sex or on the other (matrilineal and patrilineal) and which merges the direct and lateral lines but only on the side of the mother or on the side of the father, bilineal systems where each individual belongs to two kin groups defined by the merging of the direct and lateral lines on both sides, undifferentiated systems where the two lines are distinguished on both sides, then transforming radically the very concept of unit of kinship, since now it will not be definable as a stable group but rather as a changeable and decreasing intensive one going from the close relatives to the more distant ones. If I can relate these variations to other variations, for instance changes in the way what I call political authority or economic relations are organized, I will then redefine each group by a position in a system of transformations, that is, of correlated variations. I will then come to accept that something else defines me that what I used to think, and that something else is at stake in what I call family relations that relating children to parents and creating nuclear units. I took the example of Morgans anthropology of kinship terminologies and of its elaboration by the British school in what is called theory of descent groups just because it has the iconic aura of the beginnings. But it happens to be the case that this comparative analysis has been challenged (one may say refuted) on the basis of the very principle of comparative

knowledge (never accept as true but what can be constituted as a variant). This is what LviStrauss did in the Elementary Systems of Kinship, when he tried to show that it was a different parameter which was at stake here: not so much the way the belonging to one group is defined, as the way the marriages are organized, the idea being that the groups are constituted by the way they relate to one another through alliances, themselves being analyzed as exchanges in women. He gave some convincing reasons to believe that the best way to undo the equivocations and to redefine each term as a variant in a system of transformation which none of them is conscious of, is to think of them as different ways of organizing reciprocal gifts of women. If we follow Lvi-Strauss, we would then have to say that when we differentiate our relations to our parents, children, siblings and the like, we take part in a certain way of organizing a broader system of exchange. The convincing reasons are just the fact that it can give an account of more correlated variations, and tighten the redefinition of everything as a variant. But it is also clear that it has a more radical critical power (that is, it relativizes more of what we thought was universal), since it compels us to think that what is at stake for us in what we call family relations is something quite different from what we thought, since it is an exchange of women considered as gifts, rather than the constitution of kin groups. In other words, what it tells us about what kinship is is more different from what we thought kinship was about than what the theory of descent group lead us to think. The only reason I have to say that what kinship is in truth is nothing but a way of exchanging women is that it sticks more tightly to the variations of the forms of kinship and enables me to relativize my own categories in a more radical way. Therefore I have arrived, using an equivocal concept of kinship, which I applied indiscriminately to all the behaviours which looked like mine, to a redefinition of this very concept, to the effect of which I do not myself look like what I thought I was. Here we can see, quite clearly, the relativization procedure at work: a so-called universal is redefined by trying to reduce all the apparent identities to variants. Since Lvi-Strauss, other stronger versions have been proposed, in particular some which concluded that what is at stake has not even anything to do with the organization of what we call society or human relations, but rather with the metaphysical construction of the relations between humans and non-humans. I will not enter into this. But I will simply say that the fact that comparative knowledge advances by climbing over its own ruins is neither tragic nor specific to it: it is on the contrary evidence that we can stick to the relativist postulate and still be in possession of both refutability criteria and instruments of discovery which define the dynamics of any kind of knowledge: comparative knowledge is a kind of

knowledge in the sense that it enables us to reassess our own views from the perspective of newly created ones. As Descartes would say, I think this provides with everything a reasonable mind should look for: I have been brought to discover a new differential feature and to requalify things according to this new one; certain aspects of the my life (and of others life) which looked inessential to the point of being even invisible (for instance the relation of fusion, divergence or distinction between the direct and lateral line, or, in Lvi-Strauss, the relation between the husband and the brother-in-law) are revealed as being more important than the ones I would have spontaneously devoted my attention to (e.g. the relations between parents and the child) and as bearing the key of any kinship system, so that some appearances come to the fore and other retreat in the background as less important; and, finally, I dont only recharacterize myself along lines of differentiation which were unknown to me, but I also redefine what it is that we indeed have in common: for instance, kinship must not be defined by our concern with family relationships, but rather by the idea that we exchange women. This is the comparative truth about kinship. To arrive to this statement about kinship, we havent used anything else that the differences between the different ways of understanding my first equivocal concept: we dont support our claim by any sociological causal explanation or by some revelation we would have had from the brain, but simply because it is the best way to relativize the very idea of kinship, and this is what is specific about comparative anthropology. By comparing different forms of kinship, I have redefined kinship as a variant. Therefore, it is not only that I can see myself and others differently, it is also that I can define what is at stake for all of us, as it appears only from the way we equivocally differ from one another. Thus I will have redefined a pseudo-generality (kinship) as a particular anthropological concept. The same sort of endeavour could be and should be undertaken on all our generalities, like humanity, culture, thought, religion, etc. This is the nature of comparative knowledge. As you can see, it is not only about understanding other forms of life; it is also about redefining ourselves in relation to them.

2.4. Of the same and the different: logic of comparative knowledge: I may be worth coming back on the different steps of this reasoning and generalizing it beyond the example. The possibility of this self-relativization process depends on two things: first the possibility of experiencing within my categorical grid another categorical grid (which, with caution, we can call translation); next, the capacity of situating precisely these

two in relation to one another (or to more possibilities only defined in relation to one another), which is the comparison proper. I have distinguished 4 steps. 1) The first step is the step of the family resemblances. I start with a list of similarities and dissimilarities between different conducts, as they appear given my categorical grid (which is nothing but a way of assessing the similarities and dissimilarities): for instance the word mother is used to designate the mother who gives birth to a child but the child bears the name of the mother, not of the father. (This moment corresponds to the idea of the wittgensteinian anthropologist that the wood sellers pay.) If I wanted to reformulate it in terms of worlds, I could say that in the simplest case: in their world, x exist, y doesnt exist. I use my own comparative grids to assess the similarities and dissimilarities. (If I wanted to refine this concept of world, I could say that I dont simply have binary possibilities, x being perceptible or not there, but variations of degree, x being more or less perceptible, that is, more or less similar to itself. I would then say for instance that x is very apparent in this world in the sense that the value of its identity to itself is high, but y is dissimilar in the sense that its identity to itself in low, and maybe absolutely inapparent if its value is minimal. I am thinking here of the conceptual apparatus recently provided by Badiou in Logic of the Worlds.) 2) The second step is the step of the equivocations or the awareness of the misunderstandings. I observe that the conduct which looked similar to mine (designating the mother etc.) extends to conducts which I perceive as distinct from it: for instance, the word mother is used to designate both the mother, the aunt, etc.. I then start to write a list of equivocations: x = (x,y,w) z = (z,u,v), etc. 3) Thirdly, I try to characterize a variation which could account in one stroke for all these equivocations: for instance, it is for the same reason that the Iroquois dont call mother only the mother (as we do), but also the maternal aunt and others (while we dont), and that they call aunt the paternal aunt (as we do), but not the maternal aunt: the reason, as far as I can perceive it, given both my evidence and my imagination, is that they merge the lateral line with the direct line while we dont and that this variation is enough to account for all the other ones. This step is the step of the comparative hypothesis. What it means is that I try to construct within my own system of categories a variable which has a denumerable number of values, either simply because it is just a binary oppositional feature (like to merge or not to merge the direct and the lateral line), or because it has a fully ordered or systematic

set of values (like a combinatorial table or the set of natural numbers), the important point here being firstly that the field of options seems to me exhaustive (in Kants terms the list is not rhapsodic but systematic) and secondly that I can derive from each particular value the list of equivocations. Said in an informal and inductive way, I have statements of the following form: if x=(x,y,w) and z=(z,u,v), it is because V is V + instead of V , which can be reformulated in the following way: (V x=x, y=y, ) (V + x=(x,y,w) z=(z,u,v) ). In other words, I am saying that, passing a certain threshold defined by V, the identities are reassessed. I may end up having many of these principles of variation (V, U, etc.). I call this a list of comparative distinctive features. (Remark that the perception of the very coherence of my world results from the comparison itself: the systematicity of my own world can only be reached from the systematicity of the variations between the worlds. This is an important feature of the kind of holism which is at work here.) (Remark: The extent of the difference between the similarities and dissimilarities in the first step and the distinctive features of the third step give you a measure of the critical power of a comparative hypothesis; the number of equivocations which can be deduced from one particular distinctive feature (i.e. systematic difference) gives us a measure of the explicative power of a comparative hypothesis.) 4) The fourth step consists in redefining every comparative grid itself (way of making identities and differences) by mapping it into the system of such variations, and therefore defining it only by its (differential) relations to other forms. If I call a comparative grid a world, I would say then that my world is defined firstly by a number of such distinctive features by which it can be related to other possible worlds through the continuous path of equivocations (W is V+, U-, etc.), and secondly by a position in the system of the worlds related to one another by these distinctive features and in a system of positions (a space). V+ U+ UW W VW W

This I call the system of variants. Of course, the combinatorial table above is just a very poor example of one way of speaking of a system of variants, using only binary features and the poorest combinatorial technique. We can imagine much more complex models, and the question which are the ones which are best fit to account for the diversity of our understanding of the various ways of understanding is an open question, which implies both

empirical problems and formal imagination. However, we will have to go for the one which appears to be best fitted for the purposes of comparison. As a consequence, anthropology, understood this time as the theory of ourselves as human beings in the sense of that which could be differently human, or, more speculatively, as the knowledge of the subject of knowledge as it can be inferred from the variety of the forms of knowledge (i.e. the equivocations about knowledge), is, in its highest ambition, the formal theory of variance in general which is indeed best fit to account for the variations which we can experiment. In the most general sense, we are nothing other than the very way we become. But to speak of becoming in general is not very helpful: the form of becoming is an open question and the only really interesting one. Anthropology is a formal ontology of ourselves as variants. Now, I dont doubt that there are certainly many questions and worries which arise from what I have sketched, most of them probably bearing on the notion of equivocation, or misunderstanding. I wont try to cover them now, but will simply remark in conclusion that there are many things which may prevent this procedure to succeed and that I have no a priori reason to be certain that it will succeed. A survey of some of these reasons may help to understand what is at stake in this process. One reason could be that there is simply no similarity at all between my world and the other ones. But this simply means that this world and my world are completely unconnected. This situation is perfectly imaginable. It could be the case that, for instance, there are entities which perceive only events which are one million years long: they would probably not perceive us (unless we end up interfering on a scale made of units of time worth one million years each), and reciprocally. Similarly, I happen to know that there are plenty of microorganisms in this room of which I have no perception and which will never interfere in my world, because they are neither viruses, nor bacteria, nor acarids, etc. But these cases are not very interesting, because, if ex hypothesi there is no relation whatsoever between these worlds, one cannot become the other one. Another problem would be that the worlds correspond too one another just too well: all the differences expressed in one are also expressed in the other. But this simply means that we are not confronted to two different languages, but rather to the same language in two different substances, like the Code Morse and the alphabet, or the written and the oral form of English (even though, in this case, there are differences, but this can be left aside here). The most interesting case of failure is when it is indeed possible to translate, or, in Roy Wagners terms, to invent, the new language, but impossible to find any comparative

hypothesis. In this latter case, I do obtain a bilingual dictionary, I do end up being able to share in the life of the others, but I dont learn anything about any of us: true, I enrich (extend) my world (or my experience), I understand new things in the sense that I can find meaningful many things which I didnt find meaningful before, but this simply means that I add new opinions to my previous opinions, or, to put it in the terms that I used at the beginning, I perceive new voices, but I dont have any truth, I am not able to use the variety of these opinions to recharacterize them so that we both appear as variants of one another: I cannot situate us in any determinate relations to one another. New worlds appear, but this doesnt enable me to reduce them to their essence, that is, to their essence as variants: they are just juxtaposed to one another and I have no reason to prefer one opinion or one language to another: I fall back in my first condition, the condition of doubt, from which I cannot get out using only comparative tools. But, if the comparative hypothesis works, even momentarily, even precariously, and even by opening up to new problems and new more powerful solutions, then I have good relativist reasons to choose to see our worlds in conformity with what the comparison teaches me. A new appearance, a new way of looking at my world and at the relations between the worlds has made its appearance, and I am justified in thinking that this view is to be preferred to other ones, precisely because it only results from the postulate that nothing is true other than the variety of assessments of the true and the false. Therefore, if the worry was that, by accepting the diversity of opinions, we would be deprived of any procedure to create and select new opinions, we see on the contrary that to refuse to believe in anything stable or invariant in all possible worlds, we gain a new and very demanding technique for working out some truth about our world. Thanks to the comparison, I learn many things, first about myself: I learn what holds the appearances of my world together, I learn that some aspects of my world are more important than other ones because they bear the trait which is capable of undoing the coherence of my world, which we could call the critical elements); and then about what can be said of all the worlds which my worlds could become, on some particular aspects (like kinship, religion, and the like), or even, ultimately, and speculatively, on what defines a world in general: the form of a world in general is nothing that an a priori speculation can teach us, because it is only what constitutes it in such a way that it can become different. Thanks to the others truths, I find out many new truths. This, of course, is just a way of saying that contingency (in the sense of the possibility of being otherwise) is the ultimate truth of everything, including truth. We wont learn from the others anything other than how contingent we are, or, if I may say, the contingency that

we are. But this awareness of contingency is empty and probably artificial as long as we dont experience the relativization of the necessities we may share in and their redefinition in relation to one another. Lvi-Strauss concluded his Introduction to the work of Marcel Mauss, by a quote from the late master, saying that there are many dead, pale or obscure moons in the firmament of reason. I would only to the quote of this quote that, what I take to be LviStrauss intuition, is that Reason is nothing but the tireless attempt at looking at itself alternatively in the changing light of all these quivering stars. This odd astronomy may be what is left of the very project of the Enlightenment.