UNDERSTANDING OTHERS: CULTURAL ANTHROPOLOGY WITH COLLINGWOOD AND QUINE I am very grateful to Constantine Sandis and to Paul Roth in particular for commenting on earlier drafts of this paper. Without their contribution this paper would contain several more errors than those which have undoubtedly survived. Abstract: On one meaning of the term “historicism” to be a historicist is to be committed to the claim that the human sciences have a methodology of their own that is distinct in kind and not only in degree from that of the natural sciences. In this sense of the term Collingwood certainly was a historicist, for he defended the view that history is an autonomous discipline with a distinctive method and subject matter against the claim for methodological unity in the sciences. On another interpretation historicism is a relativist way of thinking which denies the possibility of universal and fundamental interpretations of historical or cultural phenomena. On these two meanings of the term historicism see Georg G. Iggers, “Historicism: The History and Meaning of the Term”, Journal of the History of Ideas 56/1, 1995, pp. 129-152. In the following I argue that at least in this second sense of “historicism” Collingwood was everything but a historicist. Quine, on the contrary, was nothing but a historicist. The goal of the comparison, however, is not to establish just who, on this definition, was or was not a historicist, but to draw a few conclusions about what a commitment to or rejection of historicism in this sense, tells us about the nature of understanding. For many philosophers the notion of meaning is essential to the project of historical understanding and the very idea of a cultural anthropology. For to understand agents with a different mindset from that of the historian or cultural anthropologist requires grasping the significance of the action they perform and this significance cannot be captured in purely extensional terms. If a Martian anthropologist visited Earth and observed a gathering of people in St Peter’s square in Rome staring at a chimney followed by outcries of joy at the sight of white smoke, how much would he know about the crowd’s behaviour without any understanding of the Catholic faith? There is often more to an action than what meets the eye and that there is more to it becomes more glaringly clear when the actions the historian tries to make sense of belong to a different period in history, different not just in time but above all in ways of thinking. In such cases it seems that an action cannot be understood or made intelligible unless one ascribes to the agents epistemic, motivational or aesthetic premises that are different from those of the interpreter. And the very fact that the understanding of actions in history requires the ascription of propositional attitudes that are different from those of the interpreters would appear to entail that the explanation of action involves appeal to intensional notions without which actions would be conceptually indistinguishable from mere bodily movements. To be sure, the need to appeal to intensional notions in the explanation of action is not confined to the interpretation of actions which took place in a distant past. Suppose Maria waters her plants at 8.30pm, which is around the time when the sun goes down and her neighbour returns from work. What is Maria doing? Is she preventing the plants from wilting or is she trying to get her neighbour’s attention? There is no discernible difference in the bodily movements that Maria performs which can tell us what she was doing, where her doing what she did was an action and not just a bodily movement. So the need to go beyond an extensional context of explanation is not limited to the task of interpreting the actions of agents with whom we have little in common. It is essential to understanding all agents. Yet, whilst the difficulty of explaining action without appealing to intensional notions is not confined to the interpretation of agents who live in different periods of time, it is brought into relief when the interpreter is confronted with agents with very different mind sets from his own. For in such cases expectations based on observations of how people tend to act when confronted with certain circumstances will have poor predictive power. One might reasonably expect a surgeon to disinfect his hands before entering the operation theatre, but such expectations will not be met if the surgeon is living in Tudor times and has no understanding of germs. It is thus no accident that the view that the explanation of action is irreducible to the explanation of mere bodily movement was articulated and defended in the context of a discussion about the nature of historical understanding and the task of a cultural anthropology. For where the epistemic and motivational premises which inform actions are not fairly uniform, the prospects for an experimental science of human nature do not appear very bright. The expectation that the appearance of white smoke will be followed by cries of jubilation will no doubt be betrayed if the crowd the anthropologist is observing is not Roman Catholic. Knowledge of purposes and background beliefs is a prerequisite for making any good predictions about what behaviour to expect. And whilst this is brought into relief in the case of agents living in a distant past, it is a feature of all action explanation, whether past or present. The view that I have just outlined was developed by Collingwood Collingwood, R.G., The Idea of History, Oxford: Clarendon press, 1944; revised edition, with an introduction by Jan Van der Dussen, Oxford, Oxford University press, 1993. who rejected the Humean claim that it is possible to extend the experimental method of observation and inductive generalization to the study of action. Collingwood’s critique of the Humean claim is to be found in his account of “scissors and paste” history in part V of the Idea of History, Epilegomena § 3 (iv), p. 257. On this see D’Oro, G., “Collingwood’s Critique of Scissors-and-Paste History revisited in the Light of his Conception of Metaphysics”, International Studies in Philosophy, 32:4, 2000, pp. 23-45. The reasons underlying this rejection, as we shall see later, had to do with the view that actions have to be understood in an intensional rather than extensional context, or, as Collingwood put it, in the “context of thought”. Yet, although the view that an action cannot be explained in purely observational terms has prima facie plausibility, it has come increasingly under pressure because it is seen to imply a commitment to intensional notions that is incompatible with a naturalistic conception of reality. As a result various attempts have been made to rethink the nature of cultural anthropology in such a way as to avoid any reference to “suspicious” semantic notions. These attempts have been articulated in the context of the philosophy of language rather than that of the philosophy of history and social science but they engage the very same question, namely, “how do we understand others?” The main challenge to the view articulated by Collingwood has come from Quine, Quine, W.V.O., Word and Object, Cambridge Mass., MIT, 1960 and Pursuit of Truth, Cambridge Mass., Harvard University Press 1990. who has endeavoured to develop an account of how we understand others without appealing to intensional notions by arguing that such a challenge is precisely that which is faced by a cultural anthropologist who has to translate the linguistic utterances of an alien tribe without having any prior knowledge of their language. Such translation is said to be radical precisely because it must proceed purely on the basis of linguistic behaviour and presuppose no knowledge of intensional notions. If it is possible for a cultural anthropologist to produce a translation manual without any native speaker’s insight into the language of the natives, so it must be possible to develop an empirical science of man and there is no obstacle in principle which prevents the anthropologist from Mars from understanding the behaviour of the crowd in Saint Peter’s square without any knowledge of the Catholic faith. Quine’s account of radical translation and Collingwood’s account of historical understanding are thus at loggerheads since the former denies what the latter asserts, namely that actions can be understood in a purely extensional context. In fact, according to Collingwood, to do away with intensional notions ultimately entails losing sight of the distinction between actions and bodily movements. Although in the latter half of the twentieth century the debate about the nature of understanding has occurred primarily in the context of the philosophy of language rather than that of the philosophy of history and social science, the concerns addressed are very close, namely whether or not intensional notions are essential or dispensable to the interpretative process. Quine argued that they are not and sought to develop an account of the nature of cultural anthropology in the light of this denial. In the following I will consider how the very idea of a cultural anthropology is influenced by certain assumptions concerning the viability or otherwise of the notion of meaning and the related concept of identity or sameness of meaning and argue that such intensional notions are only given be abandoned at considerable cost. Cultural anthropology with Collingwood The Idea of History was premised on the claim that the study of history could not be undertaken in the manner of an empirical science of man based on observation and inductive generalization. Because human beliefs and desires cannot be regarded as constant throughout different periods of time, the principle of the uniformity of nature cannot be extended to human nature and the understanding of actions in history cannot avail itself of the method employed in the natural sciences. If and when the method of observation and inductive generalization is applied to the study of actions it yields very poor results because the ability to discern similarities amongst actions for the sake of classifying them into types relies upon a prior understanding of thought processes which are not part of the extensional context available to the empirical scientist: Types of behaviour do, no doubt, recur, so long as minds of the same kind are placed in the same kinds of situations. The behaviour-patterns characteristic of a feudal baron were no doubt fairly constant so long as there were feudal barons living in a feudal society, But they will be sought in vain (except by an enquirer content with the loosest and most fanciful analogies) in a world whose social structure is of another kind... a positive science of mind will, no doubt, be able to establish uniformities and recurrences, but it can have no guarantee that the laws it establishes will hold good beyond the historical period from which its facts are drawn. Collingwood, R. G., The Idea of History, p. 223-4. The point is that generalizations aimed at predicting behaviour are reasonably accurate only when they are based on what we might call “prospective rationalizations” or practical arguments that are future-directed. Yet, predictions based on prospective rationalizations or future-directed practical arguments presuppose an appeal to an intensional context; and it is the understanding of this intensional context that gives them predictive power. The prediction that Maria will walk into her garden with a watering can will be more accurate if one understands the point of Maria’s action. If the point of her action is to attract her neighbour’s attention, then one should reasonably expect she will continue to water the plants at 8.30pm even when, with the onset of autumn, the sun will go down on the horizon much earlier than that. Since there is no one-to-one correspondence between actions and bodily movements, classifications which make use of purely extensional criteria will result, as Collingwood says, in the “loosest and most fanciful analogies”. Let me explain this further. When it comes to understanding actions the intensional context of explanation is necessary in order to determine what kind of action has been performed. To illustrate: to classify actions into, for example, “arson” requires a prior understanding of the goal with which they were undertaken. To say that an action belongs to the same class as another, for example, to say that they are both instances of murder rather than manslaughter, requires identifying them in a context that is non-extensional; it is precisely this non-extensional context that enables the action to be appropriately classified. Since there is no one-to-one correspondence between the intensional and the extensional context of explanation, one cannot take the classification of bodily movements to be a guide to the classification of actions: not all “pressing of buttons” will be “opening of windows” and not all striking of matches will be instances of arson. One and the same bodily movement - opening a window - could be compatible with different actions being performed such as, for example, letting air in or letting a fly out. And conversely, one and the same action (opening a window) could be carried out by very different bodily movements, such as pressing a button or turning a handle. Classifications of actions are thus very different from classifications in the natural sciences. For example, when Aristotle categorised virtuous actions into generous, courageous and so on, he was not providing an empirical taxonomy in the manner of the botanist or the zoologist. Courageous actions are not grouped together in accordance with certain common bodily movements, such as carrying a hand grenade or walking on a tight rope. A zoologist groups together cows and goats into mammals, and lions and leopards into felines, by noticing certain outward similarities. But the classification of actions into actions of the same kind does not appeal to perceived similarities in the bodily movements performed, but to similarities in goal. The description of actions is connected to their telos in such a way that any alteration of the goal of action entails an alteration of the action itself. Comparing and codifying actions for the sake of categorizing them into types thus requires understanding them in a teleological context. This is not to say that all actions are completely unique or sui generis and that no generalizations can be made about them, but that such generalizations are of a very different kind. Collingwood’s commitment to the view that the context of historical explanation is intensional led him to formulate a very distinctive criterion of historical understanding according to which to understand an action historically is to re-enact the very same thought that the historical agent had. He put forward his account of historical understanding not in answer to a sceptical/epistemological question: “how do we know whether we have correctly interpreted the action of an historical agent?”, but in answer to a conceptual question: “what is historical understanding?” He answered the conceptual, as opposed to the epistemological question, by saying that an action is understood historically when the interpreter rethinks the very same thought as the historical agent. Re-thinking is normative or criteriological activity, because thoughts are ordered not temporally but logically. When the historian re-enacts the thought of an agent he/she establishes an internal or non empirical connection between the explanans and the explanandum. Thus, for example, the claim that Oedipus murdered his father because he wanted to marry his mother is a re-enactive explanation because it explains the action in relation to a purpose or goal that makes sense of it. Had Oedipus’s goal been different then his action would not be parricide, for what determines the description of his action as parricide rather than either murder or manslaughter is the internal relation with which it stands to its goal. This is what it means to say by the claim that there is an internal or conceptual connection between the explanans and the explanandum. If the historian selects a different explanans (as in “alter the presumed goal”), this affects the description of the action. Collingwood expressed the claim that re-enactive explanations establish an internal, non-empirical connection between the explanans and the explanandum by saying that “when he [the historian] knows what happened, he already knows why it happened.” Collingwood, R.G., The Idea of History, Epilegomena § I, “Human Nature and Human History”, p. 214. The claim has often been viewed as paradoxical but it is in fact making the less controversial point that since the connection between the explanans the explanandum is conceptual, describing Oedipus’ action as parricide entails that the goal of Oedipus’ action was to kill his father king of Thebes, not to kill a stranger at the crossroads. Now this particular historical explanation happens to be false for, as we all know, Oedipus only willingly killed a man at the crossroads and so was guilty of murder, not parricide. Yet, although the explanation is false, it has the correct logical form, that is the logical form that Collingwood believes is required in order to explain something historically, as an action rather than a mere bodily movement. Were the explanation correct in content as well as in form, then the interpreter would have the same thought as Oedipus, not a thought of a similar kind. To get the explanation right in its form is to establish an internal conceptual connection; to get the explanation right in content is to have the same thought as the historical agent. Many action explanations may be correct in form but not in content. In this case they will not have successfully identified the thought of the agent. But all action explanations (and for Collingwood to say that an explanation is historical and that it has actions as its explanandum is one and the same) are concerned with tracing internal relations, and this is why they differ from explanations in the natural sciences. In so far as re-enactment provides a criterion of understanding it explains what understanding is, not how it is acquired. Re-enactment, in other words, does not tell us that “identity of thought” has been achieved but what historical understanding yields when it is successful. Compare re-enactment as a criterion of understanding with a criterion for telling the difference between waking and dreaming experiences, such as the one which states that whereas dreaming experiences are chaotic, waking experiences are orderly and law-like. To be in possession of this criterion is to be able to make the distinction between dreaming and waking experiences but not necessarily to be right in each single identification of a waking experience as a waking experience. Similarly, for Collingwood, re-enactment is a criterion of historical understanding in the sense that it enables us to make a distinction between explanations which establish a conceptual connection and explanations which make an empirical connection, i.e. between pseudo-historical explanations and historical explanations proper. When the criterion is successfully applied then the thought of the historian and that of the agent are one and the same. So if the historian does get it right he does not create a numerically distinct replica of the original thought since, if the thought that the historian has is qualitatively identical to that of the agent, there is only one thought. Much in the way in which the English, Italian and French expressions: “I love you”, “ti amo”, and “je t’aime” mean one and the same thing, and translating successfully from one language to another requires taking stock of the fact that they mean one and the same thing, so understanding the action of an historical agent involves rethinking the same thought as the historical agent. The historian, like the translator could of course get it wrong. But this is not the point for, although the historian, investigative journalist/biographer can get it wrong it is in principle possible to get it right, because it is in principle possible to have the same thought as the historical agent. There are, therefore, criteria for telling the difference between what would count as a correct and what would count as incorrect understanding. A correct explanation is one which succeeds in achieving identity of thought. An incorrect explanation is one which fails in this task. Many have found the claim that to understand the action of an historical agent is to have the very same thought of the historical agent puzzling. For, it has been asked, how can two people share one and the same thought? Yet the claim is puzzling only if one conflates qualitative with numerical criteria of identity. Two bodily movements are of the same kind if they share certain observable qualities (for example they are both raisings of the arm). The fact that they belong to the same kind does not imply that they are numerically one and the same. But how many thoughts there are depends precisely on the nature of the thought, for two thoughts are numerically distinct only in so far as they are qualitatively different. Talk of sameness or identity of thought may appear strange at first, but to say that when a historian understands an agent he has the same thought as that agent, not a thought of the same kind, is in fact no stranger than saying that there is only one masterpiece entitled Meditations on First Philosophy which was written by Descartes, not as many manuscripts as there are printed copies of that masterpiece. When one denies that it is appropriate to speak of the printed copies of the Cartesian Meditations as instances of the Meditations, in the way in which one might speak of Dolly as an instance of a sheep, one effectively denies the wisdom of applying numerical criteria of identity in the context of thought. Qualitative identity is the goal of understanding; it is what the historian achieves if and when he gets it right. On Collingwood’s account of re-enactment see, D’Oro, G., “Collingwood on Re-enactment and the Identity of Thought”, Journal of the History of Philosophy, 38:1 (2000): 87-101; Saari, H., ‘R. G. Collingwood on the Identity of Thought’, Dialogue: Canadian Philosophical Review 28 (1989): 77-89 and Van der Dussen, J., “The Philosophical Context of Collingwood’s Re-enactment Theory”, International Studies in Philosophy, 27/2 (1995): 81-99. What however might make the historian think that he has got it right? Collingwood asserts both that when the historian re-enacts the thought of an historical agent the historian and the agent have the same thought and denies that his account of re-enactment is an instance of the copy theory of knowledge. On Collingwood’s discussion of re-enactment and the rejection of the copy-theory of knowledge see The Idea of History, Epilegomena, § 4. These claims seem to jar since it may be natural to assume that in order to have the same thought as the agent the historian must form in his mind an exact copy of the thought the historical agent has. And this in turn implies that the agent consciously recited a thought in his head. But whilst it may be natural to be drawn to this interpretation this is in fact a distortion of Collingwood’s account of re-enactment. For re-enactment, as we have seen, consists in the tracing of internal relations and as such it is a normative or criteriological activity, not a descriptive one. If the historian concludes that he has re-enacted the thought of the historical agent, it is not because he has made psychic contact with the ghost of the agent and found a match, but because the net of internal relations he has established is so tight and persuasive that it is difficult for him to conceive that the agent might have acted for different reasons. When the historian gets to the point where he can conceive of no better re-enactive explanation as to why the agent should have acted in that way, then he concludes that he has re-enacted the same thought as the agent. But such a claim is not achieved by magical or supernatural means. It is the conclusion of an inferential process that as Collingwood says, might begin with the “outside” of the action, but cannot end there. At times Collingwood compares the method of history to that of a detective who examines the evidence to solve a murder case. Collingwood, R.G., The Idea of History, § 3 Historical Inference, especially pp. 266-68 “Who killed John Doe?” The detective will be presented with evidence, reports and other information that requires connecting. When the evidence leads incontrovertibly to a conclusion, the detective will know that he has found the murderer. Such a conclusion, Collingwood claims is not known with probabilistic force, in the manner in which inductive inferences lend probabilistic support to a conclusion. The force of the detective’s argument will be so strong as to compel more in the manner of a deductive than a probabilistic inference. Understood in this way re-enactment does not entail that the historical agent must have consciously recited a thought in his head. To interpret the cheering of the crowd in St Peter’s square the historian need not presuppose that the members of that crowd recited a thought of this kind “when smoke comes out of the chimney a new Pope is elected, I am cheering because I am happy a new Pope has been elected”. The identity of thought as a criterion of historical understanding, requires the much weaker claim that those cheering in St Peter’s square could recognize what the historian proffers an explanation of their action as reflecting a thought process that they could have consciously articulated, not one that they did. Since re-enactment is the result of an inferential process, not a description of a psychological one, Collingwood’s defence of the claim that action must be explained in an intensional context is not premised on what Ryle called the Cartesian myth of the ghost in the machine, Ryle, G., The Concept of Mind, London: Penguin Books, 1990, pp. 17ff. First published by Hutchinson, in 1949. or what Quine refers to as the museum myth of meaning. Quine, W. V. “Epistemology Naturalized” in Ontological Relativity and Other Essays, New York: Columbia University Press, 1969, p. 27. Collingwood’s commitment to the claim the context of action explanation is intensional rather than extensional is motivated not by a commitment to the view that thoughts are inner/psychological states and to re-enact them is to describe the thoughts of the agents. It is the result of the view that actions have to be understood in a normative context. The goal of the human sciences, for Collingwood, is not to represent but to make sense. Re-enactment does not, therefore, presuppose what one might refer to as meaning realism. See Roth, P., Meaning and Method in the Social Sciences, chapter 6, “The Myth of Meaning Realism”. What it does presuppose is that the context of action explanation is normative and that understanding others is more like solving a puzzle than cataloguing plant or animal species. It is this assumption rather than the commitment to meaning realism that, as we shall see, constitutes the bone of contention between his and Quine’s account of how we understand others. Collingwood developed the claim that it is not possible to explain actions by using the method of observation and inductive generalization in the context of his philosophy of history. But his claim was not simply that the actions of past agents cannot be understood by the method of observation and inductive generalization, but rather that no action, whether past or present, can be understood in a purely extensional context of explanation: If it is by historical thinking that we re-think and so re-discover the thought of Hammurabi or Solon, it is in the same way that we rediscover the thought of a friend who writes us a letter, or a stranger who crosses the street. Collingwood, R.G., The Idea of History, p. 218. Understanding is historical not because it is concerned with the actions of past agents but because all actions, whether past or present, must be understood in the context of thought. The need to appeal to an intensional context of explanation is brought into relief in the case of agents belonging to different cultures who do not share the same beliefs and desires as the interpreter, but the actions of past agents are only a special case of actions in general and actions in general have to be understood, as he put it, in a thought context. He expressed this claim a little controversially by saying that actions have an inside or thought side that events lack, thereby opening the way to a flood of criticisms aimed at discrediting his view of the action/event distinction as relying on a Cartesian theory of the mind. But the inside/outside distinction is in fact just a highly metaphorical way of capturing the claim that the explanation of action is a species of justification and that the method of the human sciences is different from that of the natural sciences, which rely on observation and inductive generalization. W.H. Dray built on Collingwood’s claim that re-enactment is a normative activity and mobilized this argument against Hempel’s claim that the structure of explanation, both in the human and natural sciences, is nomological. Re-enactment as a criterion of historical understanding entails that it is at least in principle possible for a historian to transcend his own world view and entertain the very same thought as an agent who lived in a distant past. This of course is not to say it could always be done or that it is possible to know whether the interpreter has succeeded with absolute certainty. But it is to say that it would be in principle possible to do so. The reasons why it is in principle possible to entertain the same thought as an historical agent is that the criteria of identity which apply in an intensional context or the context of thought are qualitative, not numerical. If Paul and James raise their arm there will be two numerically distinct bodily movements. But if two agents recite the same thought there will be only one thought, not two thoughts of the same kind, for thoughts are distinguished numerically from one another only in so far as they are qualitatively distinct. Thus for example if you and I recite the syllogism: “All men are mortal, Socrates is a man; therefore Socrates is mortal” there is by Collingwood’s lights, only one thought, not two thoughts of the same kind. By the same token if the historian re-enacts the same practical syllogism as another agent, such as “I would like my plants to thrive, in order to thrive plants need to be watered; therefore I water my plants”, the historian has the same thought as the agent, not a thought of the same kind. On Collingwood’s anti-relativistic conclusions about the recovery of authorial intentions see Kobayashi, C and Marion, M., “Gadamer and Collingwood on Temporal Distance and Understanding”, History and Theory 50/4, 2011, pp 81-103. There is nothing historicist about this account of the nature of historical understanding. If by historicism one means the claim that it is impossible to provide fundamental and universal interpretations, then Collingwood is anything but a historicist. His account of historical understanding entails precisely that it is at least in principle possible for two persons, be they contemporaries or not, to have the same thought and that such thought need not be altered by the mind-set of the interpreter, provided of course that due care is taken in suspending one’s own beliefs. Whilst we might reasonably gasp at the Tudors’ lack of hygiene and, in an hypothetical “Horrible Histories” scenario where we were offered the services of a Tudor surgeon we would be well advised to decline them, there is no barrier in principle to viewing the world from their perspective and understanding why, in their own times, Tudor surgeons were not regarded to be negligent and morally culpable for the high mortality rate amongst their patients. The differences in Tudor and twentieth century beliefs about what cures are causally efficacious provides no conceptual barrier to understanding why the average Tudor person would have trusted their lives in the hands of Tudor surgeons. But we could not understand why any sane person could possibly accept the services of a Tudor surgeon unless we explain their action in what Collingwood called the “context of thought”. Independently of such an intensional context of explanation their action would remain at best unexplained and at worse unintelligible. Collingwood’s endorsement of historicism in the first sense of the term, as the claim that the human sciences have a distinctive method and subject matter that differs from that of the natural sciences, thus leads to the rejection of historicism in the second sense, as the denial of the view that we can determine the meaning of action from the point of view of historical agents and that the interpretations of the thoughts and actions of past agents need not, like scientific hypotheses, be revised in line with the progress of natural science. It is in so far as the historian can reconstruct inferential patterns that he no longer shares, such as the view that one ought not disturb devils in case they will retaliate, that the historian can see the world from the point of view of historical agents. Considering the appropriate attitude a historian should have towards agents who believe in devils and reason from this epistemic premises that they ought not to cross a mountain chain because it is inhabited by evil spirits, Collingwood claims that wrong ways of thinking are just as much historical facts as right ones, and, no less than they, determine the situation (always a thought situation) in which the man who shares them is placed. The hardness of the fact consists in the man’s inability to think of his situation otherwise. The compulsion which the devil haunted mountains exercise on the man who would cross them consists in the fact that he cannot help believing in the devils. Sheer superstition, no doubt: but this superstition is a fact, and the crucial fact in the situation we are considering. Collingwood, R.G., The Idea of History, p. 317 (IH, 317) The point that Collingwood is making here is neither that the historian has to agree with the agent (it is not whether the agents’ beliefs track the truth that is at stake), nor that there is a law-like generalization according to which agents who believe in devils will act in certain ways (the explanation of the agent’s behaviour is not causal/nomological) but that there is an internal connection between the content of the agent’s beliefs and his or her action and that to understand the agent is to expose this conceptual/internal connection. And it is because the connection that the historian uncovers is an internal conceptual connection that its status is not sensitive to the passage of time, much as the claim that vixens are female foxes cannot be falsified by future empirical evidence. Future empirical evidence may show that there are no devils but such a discovery has no bearing on the claim that if one believed in them one ought to be wary of them. But of course all of this holds only if one accepts the analytic/synthetic distinction. Cultural anthropology with Quine Collingwood’s account of historical explanations was premised on the view that there is such a thing as getting it right when attempting to explain the actions of others. When the historian, biographer/investigative journalist get it right they achieve identity of thought. As we have seen, the historian/biographer/investigative journalist may get it wrong in this or that occasion, but this is not the point, for they could also get it right. The view that it is possible to get it right, even in principle, has been challenged by Quine, who has defended the thesis of the indeterminacy of translation. Quine, WVO, Word and Object, Cambridge Massachusetts, the MIT Press, 1960, chapter 2; Pursuit of Truth, Cambridge Massachusetts, Harward University Press, 1990 part III. He articulated his claim in the context of the philosophy of language rather than that of the philosophy of history and social science but the target of his claim that “there is no fact of the matter” Quine, WVO, “Things and Their Place in Theories” in his Theories and things, Cambridge Massachusetts, Harvard University Press, 1981, p. 23. about the correct translation is arguably precisely Collingwood’s claim that there is a criterion of understanding that is typical of the human sciences. Quine illustrates the thesis of the indeterminacy of translation by imagining a “radical scenario”, a scenario in which a field linguist is confronted with a completely alien culture, so alien that the only explanatory tools available to the cultural anthropologist would be those of the natural sciences and the behaviour of the agents under observation could only be accounted for in the manner of natural phenomena such hurricanes and volcanic explosions. As Hylton puts it, “the point of the general idea of radical translation is to give an approach to language which is evidently empirical, to see how much can be made of the idea of meaning.” Hylton, P., Quine, New York and London, Routledge, 2007, p. 199 The field linguist observes a native utter the word “gavagai” in the presence of a rabbit and writes down “rabbit” in his translation manual. He codifies linguistic behaviour in the manner in which a botanist catalogues plants and a zoologist classifies animals. Much as the anthropologist from Mars observing earthlings cheering at the sight of white smoke emanating from a chimney in Rome, the field linguist can only appeal to behavioural data. The behavioural evidence, Quine argues, is insufficient to determine what the native means when he utters the word “gavagai”, whether, for example the native means a whole rabbit or undetached rabbit parts. Since, Quine argues, the only evidence for imputing meanings to others is based on observations of their behaviour and behavioural evidence cannot legitimate the ascription of a particular set of intensions over another, translation is indeterminate. Quine calls this translation scenario “radical translation” precisely because the only epistemological tools available to the translator are those drawn from the empirical sciences. The thesis of the indeterminacy of translation is a very strong one because Quine is not claiming that there is a correct translation, although it is very hard to know what this is. As such the claim concerning the indeterminacy of translation is to be distinguished from Quine’s claim concerning the underdetermination of truth by evidence. On this see Roth, P., Meaning and Method in the Social Sciences, p. 136, chapter 6; Moore, A., The Evolution of Modern Metaphysics, chapter 12, p. 317. See also Hylton, P., Quine, New York and London, Routledge 2007, p. 201-202. There is no fact of the matter about translation not because the linguist, just like the historian/investigative journalist/biographer could get it wrong, but because since empirical evidence systematically underdetermines translation, and since there is no other criterion applicable, there is no standard by which one translation could be deemed better than another. Two linguists, for Quine, could come up with different and incompatible translation manuals which are equally successful. Though the scenario of a field linguist confronted with a completely alien tribe provides a paradigmatic example of radical translation, it also illustrates the conditions under which all translation operates. Translation for Quine, is indeterminate whether the target language is a remote or obscure language (Jungle) or French. We cannot say that the expressions “Je t’aime”, “I love you” or “ti amo” are equivalent translations because they are synonymous since the idea of identity of meaning or thought does not belong to the conceptual repertoire available to the empirical scientist whose job is to establish purely external relations. Collingwood and Quine are therefore similarly radical in their claims. For Collingwood it is by historical understanding that we explain not only the actions of agents who lived in a distant past, but also those of our contemporaries. The claim that actions, in order to be understood, must be understood as an expression of thought, is a claim about the nature of action, whether present or past, and how actions differ from events. Equally for Quine, the claim that translation must be radical applies not only to the field linguist but also to the contemporary translator. On this see Glock, H.J., Quine and Davidson on Language, Thought and Reality, Cambridge: Cambridge University Press, 2003, chapter 6. Since all translation is radical, all translation, whether from a modern or ancient language, is indeterminate. What justifies Quine’s strong conclusion that there is no such thing as the correct translation? This robust conclusion is not warranted by the consideration that in a difficult scenario, such as the one that confronts the field linguist translating from Jungle to English, it may be hard to get it right because the thought patterns of the agents are very different from those of the linguist. Collingwood, for example, conceded that some periods of history are hard to fathom, but did not conclude from this that historical explanation must be radical in the sense that it must proceed by the methods of the natural sciences. He explicitly denied that the rational inaccessibility of certain periods of history warrants shifting from one kind of enquiry to another. The fact that there are periods of history the historian cannot comprehend does not entail that historians should study the past in the manner in which a palaeontologist studies fossils or a mineralogist studies rocks: “Certain historians, sometimes whole generations of historians, find in certain periods of history nothing intelligible, and call them ‘dark ages’; but such phrases tell us nothing about those ages themselves; though they tell us a great deal about the persons who use them, namely that they are unable to re-think the thoughts which were fundamental to their life.” Collingwood, R.G., The Idea of History, p. 218-9. The inability of the historian to discover intelligibility in some circumstances is, for Collingwood, insufficient to warrant the general conclusion that re-enactive explanations can never succeed and the historian would do better abandoning the hermeneutic method in favour of an empirical one. Just as the consideration that occasionally, if the lighting conditions are poor, we may not be able to determine whether what we are seeing is a black cat or a dustbin liner does not warrant the conclusion that we can never determine whether what we see is or is not a cat, so the consideration that some historical periods are opaque to the historian does not entail that every past stage of human history should be investigated as if it were a natural phenomenon. But if the consideration that determinacy of meaning is particularly difficult to achieve in some circumstances does not entail that it cannot be achieved in any circumstances, what is it that warrants Quine’s claim that translation is indeterminate whether the target language is Jungle or French? The thesis of the indeterminacy of translation is the logical consequence of trying to do cultural anthropology against the background of the rejection of the analytic/synthetic distinction Quine, W.V.O., “Two Dogmas of Empiricism”, Philosophical Review, 1951, 60 pp. 20–43. as an unempirical dogma of empiricism. Hume had claimed that the building blocks of knowledge have their source in experience. But the Humean fork still allowed for the possibility that not all knowledge is grounded in sense experience, since he allowed relations of ideas to slip through the principle of empiricism. And the same is true of logical positivists who interpreted the empiricist principle in verificationist rather than genetic terms, but still made an exception for analytic statements. In the attempt to achieve a form of empiricism that is pure and stripped of any extraneous rationalist overtones Quine argued that no statement has its own meaning which is true or false independently of reality since all statements face the tribunal of experience, not individually but as a body. As a consequence of Quine’s holism analytic statements become subject to the same verification and falsification conditions as empirical ones. Whilst some statements may be more resistant to falsification through the course of experience, there are no statements which are completely impervious to it and which are true come what may. Quine’s holism ushers in a form of historicism which ultimately has its roots in a commitment to a monistic conception of verification. Collingwood, alongside many non-reductivists of his generation See “From Causalism to Anti-causalism and Back” in D’Oro, G and Sandis, C. (eds.) Reasons and Causes: Causalism and Anti-causalism in the Philosophy of Action, Palgrave 2013., had endorsed the view that understanding the world of human affairs requires grasping a system of internal relations on which the passing of time has no significant bearing because the tribunal to which historical explanations are accountable is that of reason rather than experience. Explanations of actions succeed to the extent that they enable us to make sense of the deeds even if purely from the agent’s point of view, rather than that of the historian. For Quine, by contrast, there is no net of internal relations which are impervious to the passage of time precisely because there are no statements which are analytic in the sense permitted by Hume and the logical positivists. The human sciences do not have a distinctive domain of enquiry because there are no such things as internal connections which come with their own criterion of truth and falsity. Did the people cheering at the white smoke in St Peter’s square do so because they saw white smoke or because they saw a bird sitting on the chimney at the very exact time in which the smoke started to come out? If the symbolic significance of white smoke is fixed, then the historian could understand the action of those agents against that bedrock; but if the bedrock can move, any hypothesis becomes feasible, including the one that the people were cheering because a bird sat on the chimney. Once statements which are held to be true in virtue of what they mean are allowed to be revised in the light of experience, then translation must be indeterminate, not because in some cases it may be difficult to find out what the native means, whether for example, when the native utters “gavagai” he means a whole rabbit or a bundle of rabbit parts, but because once the notion of truth in virtue of meaning, and with it, the idea of identity of meaning is cast aside as incompatible with empiricism in its pure form, synonymy can no longer be advocated as a criterion of correct translation. Collingwood’s endorsement of historicism in sense 1, according to which the human sciences have a distinctive method and subject matter (the normative uncovering of internal conceptual connections), that is distinct from that of the natural sciences leads to the rejection of historicism in sense 2, according to which there is no such thing as universal interpretations which are impervious to the passage of time. By contrast, Quine’s rejection of historicism in sense 1 ushers in a commitment to a form of historicism in sense 2. Quine’s rejection of methodological pluralism in the sciences thus aligns his naturalized epistemology with a form of extreme historical relativism. There cannot be a determinate translation because determining the translation would require an appeal to an intensional context of explanation that is ruled out by Quine’s extension of the verificationist principle to all judgments, including analytic ones. Quine’s rejection of the analytic/synthetic distinction undermines, in one fell swoop, both the ontological/metaphysical claim that there are inner hidden psychological items which escape empirical verification and the methodological claim that the human sciences are hermeneutic sciences with their own standards or criteria of correctness because they have, as their subject matter, a web of internal relations. Quine’s rejection of the analytic/synthetic distinction thus paves the way to the acceptance of the very kind of historicism of which Collingwood has often been accused. As Alan Weir points out, perhaps a little too unkindly to continental philosophers, “the idea of the indeterminacy of meaning has more than a whiff of smoke-filled cafés on the banks of the Seine about it, though Quine’s arguments for the indeterminacy of translation belong firmly to the tradition of logical empiricism.” Alan Weir, “Indeterminacy of Translation” in E. Lepore and B. Smith (eds.) The Oxford Handbook of Philosophy of Language pp. 233-249. If I have not misunderstood him, Weir is suggesting that Quine is a postmodern historicist. And historicism, in the sense that there can be no right/wrong interpretation in the human sciences, results precisely from Quine’s rejection of historicism in the first sense, according to which the human sciences have a distinctive method and subject matter. Paradoxically it is not Collingwood the philosopher of history and hermenutician who is a historicist, but Quine the analytical philosopher and heir of logical positivism who in fact is. PAGE \* MERGEFORMAT 16