On Tradition
By Mark Bevir

            Tradition can be a highly evaluative concept.  Conservatives often evoke the idea of tradition to express reverence for continuity and the past.  Tradition can act as an anti-theoretical concept deployed to question the role of doctrine and reason within social life.^[i][1]  Traditions allegedly validate social practices by providing an immanent guide to how one should behave.  Any abstract doctrine or reason informing such a guide is best - or perhaps of necessity - left unarticulated since such abstractions are inherently destructive in their effects on social order.  The ability of traditions to confer legitimacy on social practices helps to explain why cultural nationalists, states, and even radical movements have tried to invigorate their political projects by inventing appropriate traditions, symbols, and rituals.^[ii][2]

            Yet whilst tradition can be an evaluative moral and political concept, it also plays a vital role as an ontological and explanatory one.  Historians often explain features of works, actions, and practices by locating them in the context of a particular tradition.  Even when scholars explicitly reject the concept of tradition, they typically adopt a related concept to indicate the importance of social and historical contexts for a proper understanding of particular works, actions, and practices.  It appears that a concept such as tradition, structure, heritage, or paradigm is integral to our understanding of the human condition.  One argument for believing this to be so - the one I will adopt - derives from meaning holism.  What is more, this argument encourages us to unpack the relationship of individuals to their social and historical contexts in a way which suggests the concept of a tradition is preferable to that of a structure or paradigm.  Finally, because the ontological and explanatory notions of tradition clearly overlap with one another, we can use the ontological concept thus derived from semantic holism to say something about, first, the idealization procedures by which historians should construct traditions to explain a particular object, and, second, the nature and limits of such explanations.

The Necessity of Tradition

            Analyses of the forms of explanation that historians should adopt with respect to works, actions, and practices typically revolve around two sets of concepts.  The first set includes concepts such as tradition, structure, and paradigm.  These concepts embody attempts both to specify how we should analyse the social context in which individuals reason and act, and to indicate how much weight we should give to the social context as a factor in their reasoning and acting.  The second set includes concepts such as anomaly, reason, and agency.  These concepts embody attempts to specify how we should analyse the processes by which beliefs and practices change, and, more especially, the role played by particular individuals in these processes.  Within both sets of concepts, there are, of course, numerous further debates over how we should unpack the relevant concepts.  Scholars debate, for example, the respective weights we should ascribe to economic and political factors within the social context, or the extent to which the unconscious, desire, and reason effect the individual performance.  Nonetheless, these two sets of concepts are clearly vital ones for a study of tradition since they concern the relationship of the individual to his social inheritance.

            There are philosophers who appear to believe that the individual is wholly autonomous, that is, able to transcend totally the influence of tradition.^[iii][3]  A faith in such autonomy often draws support, explicitly or implicitly, from a strong empiricism.  Empiricists generally argue that people arrive at webs of belief as a result of pure experiences.  This can suggest that the historian can explain why people held the beliefs they did by reference to their experiences alone: the historian needs to consider only the circumstances people find themselves in, not the ways in which they construct or interpret their circumstances through the traditions they inherit.  Yet such a strong empiricism seems highly implausible nowadays, largely because of the powerful arguments in favour of various types of meaning holism.  Here is not the place to follow the detour required for a full defence of holism.  Instead it must suffice to note that the vast majority of philosophers now accept some form of holism, and, moreover, that holism informs many of the major developments in modern philosophy, including the rejection of pure observation by philosophers of science such as Thomas Kuhn, the analysis of meaning and interpretation by philosophers of knowledge such as W. V. O. Quine and Donald Davidson, and, to a lesser extent, the analysis of intentions and beliefs by philosophers of mind such as David Lewis.^[iv][4]  

            Meaning holism shows a faith in autonomy to be mistaken.  Certainly people come to believe the things they do only in the context of their own life-histories: my beliefs are my beliefs precisely because I have come to accept them as mine.  What interests us, however, is why certain beliefs should become part of a particular life-history: why do I hold the particular beliefs I do?  Because we can not have pure experiences, we must necessarily construe our personal experiences in terms of a prior bundle of theories.  We can not arrive at beliefs through experiences unless we already have a prior set of beliefs.  Experiences can generate beliefs only where there already is a set of beliefs in terms of which to make sense of the experiences.  Thus, strong empiricism is wrong: we can not explain a belief by reference to the pure experiences of the relevant individual.  Our experiences can lead us to beliefs only because we already have access to sets of belief in the form of the traditions of our community.  Individuals necessarily arrive at their beliefs by way of their participation in traditions.^[v][5]

            Our social inheritance constitutes the necessary background to the beliefs we adopt and the actions we perform.  Some philosophers adopt a very strong version of this conclusion.  They argue that some sort of social structure - a paradigm or episteme, for example - fixes, or at least limits, not only the actions we can perform successfully but also our very beliefs and desires.^[vi][6]  Strong structuralists typically argue that meanings, beliefs, and the like are the products of the internal relations of self-sufficient languages or paradigms.  They leave little, if any, room for human agency.  Surely, however, such a strong structuralism is in error.

            Certainly people adopt their beliefs against a background of traditions that already exist as a common heritage: I come to formulate my beliefs in a world where other people already have expressed their beliefs.  What interests us, however, is how the beliefs of particular individuals relate to the traditions that they inherit: how do I develop my beliefs in relation to the beliefs other people already hold?  Here strong structuralists suggest that traditions, structures, paradigms, and the like determine, or at the very least set definite limits to, the beliefs people might adopt so the actions they might attempt.^[vii][7]  Imagine that we could indeed identify necessary limits imposed by social contexts on the beliefs individuals can adopt.  Because the social contexts would impose these limits, they could not be natural limits transcending all contexts, as is the natural limit to how fast I can run.  What is more, because we could identify these limits, we could describe them to individuals within the relevant social context, so, assuming that they could understand us, they could come to recognise these limits, and so understand the sorts of beliefs they could not adopt.  But because they could recognise these limits, and because these limits could not be natural limits, therefore, they could transcend these limits, so these limits could not really be limits at all.  Because they could understand the sorts of beliefs these limits preclude, and because there could not be any natural restriction preventing them from holding these beliefs, therefore they could adopt these beliefs, so these beliefs could not be beliefs they could not come to hold.  For example, if we could recognise that such and such a community of monarchists could not possibly form a republic because of the social context, we could explain the nature of a republic to them, so they could become republicans, and, if enough of them in sufficiently powerful positions did become republicans, they then could found a republic.

            There are two features of this argument against strong structuralism that seem to need defending.  The first is the apparent proviso that we can describe a limit to the people it effects only if we are their contemporaries.  This appears to leave open the possibility of social contexts imposing limits that we can not recognise at the time.  We recognise them only historically after they cease to operate.  However, this apparent proviso does not actually apply because the argument concerns the conceptual, not the empirical, pre-conditions of limits.  Thus, the argument can be rewritten as a thought-experiment.  If we imagine an outsider who is aware of the limit entering into the relevant context, this person could describe the limit to the relevant individuals at which point it would cease to be a limit for the reasons already given.  The fact that we envisage the limit being transcended in this thought-experiment shows that it is a contingent, not a necessary, limit; after all, if it was a necessary limit imposed by the social context, we would be able to envisage people transcending it only after the social context had changed so as to prevent it operating.  It is possible that a critic might complain that the social context changes as soon as someone who is aware of the limit arrives on the scene.  But this will not do because the critic thereby defines the social context to include people's awareness or lack of awareness of the purported limit.  Thus, the critic makes the purported limit a mere description of the facts.  He replaces the claim "people can not come to believe X because of the social context" with the claim "people can not come to believe X for so long as they do not believe X", and this latter claim is not very illuminating.

            The second feature of the argument against strong structuralism that seems to need defending is the assumption that the individuals affected by a limit could understand our account of it.  Although the possibility of translation between radically different sets of beliefs is a premise of our argument, this essay is not the place to defend it at any length since doing so would require a major detour from our main theme.  Instead it must suffice to note that we have no reason to assume that people can not translate between sets of beliefs no matter how different they are.  When the individuals concerned first approached our account of the limit, they might not have the requisite concepts to understand us, but surely they would share some concepts with us, and surely they could use these concepts as a point of entry into our worldview, so surely they eventually could come to understand us.^[viii][8]  Indeed, if they did not share some of our concepts, we would not share any of their concepts, so we would be unable to translate their beliefs into our terms, so we would have been unable to identify any limits on the beliefs they could adopt in the first place.^[ix][9]

            So, we can not identify any limits imposed by social contexts upon the beliefs individuals might come to adopt.  If we could do so, we could describe these limits to these individuals who then could transgress these limits in a way which would show they were not limits at all.  Moreover, because there is no possibility of our ever identifying a restriction imposed by a social context on the beliefs individuals can adopt or the actions they can decide to perform, we must conclude that the idea of such a restriction rests on a conceptual confusion.  Social contexts only ever influence, as opposed to decide or restrict, the nature of individuals.  This means that social contexts must be products of the undetermined agency of individuals.  Traditions, structures, and paradigms can not be self-contained systems because they depend on the beliefs and actions of individuals, and they do not decide the nature of these beliefs and actions.

            Perhaps our rejection of strong structuralism, our insistence on the fact of human agency, will seem to some critics to be incompatible with our earlier insistence on the unavoidable nature of tradition.  In fact, however, our reasons for evoking tradition allows for human agency.  The variety of agency that survives our appeal to tradition is the ability of individuals to extend and modify the traditions they inherit.  Just because individuals start out from an inherited tradition does not imply that they can not go on to adjust it.  Indeed, traditions change over time, and we can not explain these changes unless we accept that individuals are capable of altering the traditions they inherit.  The easiest way to make this point is counter-factually.  Traditions arise from the beliefs and actions of numerous individuals, so if they determined the beliefs of individuals, we would have a closed circle that would preclude change.  Imagine that the totality of the beliefs held and the actions performed by individuals in a society is as it is, so the traditions therein are as they are.  Because traditions are emergent entities, they could not alter unless this totality of beliefs and actions changed.  But if traditions really determined beliefs and actions, this totality could not alter unless the traditions did so.  In order to explain change, therefore, we must summon up individuals who can extend and modify the traditions that provide the starting points from which they arrive at their beliefs and practices.

            The possibility of agency even extends to the ability of individuals to reflect on different traditions and thus decide to migrate from one to another.  Just because individuals start out against the background of an inherited tradition does not imply that they can not end up within another one.  Indeed, people do convert from, say, Christianity to Islam, and we can explain their doing so only by accepting that  individuals are capable of crossing any boundaries allegedly dividing traditions.  Some scholars appear to cast doubt on the possibility of such boundary crossings by arguing that individuals who adhere to one tradition can not grasp the meaning of the concepts constitutive of another.^[x][10]  Doing so, however, requires them to embrace a strong thesis of incommensurability that we should reject as false.^[xi][11]  For a start, even when two traditions make use of different concepts, they still can overlap in ways which might provide people who subscribe to one with points of entry into another.  Moreover, even if two traditions did not overlap at all, the adherents of the one still could observe the practices of the adherents of the other and so learn the meaning of the concepts embodied in these practices.

            To recognise the inevitable influence of tradition on individuals is not to deny human agency.  Although individuals must begin their journey against the background of tradition, they later can modify that tradition; although they are inescapably influenced by it, they are not determined by it.  Indeed the ability to develop traditions is an essential part of our being in the world.  We are always confronting slightly novel circumstances that require us to apply tradition anew, and tradition itself can not fix the nature of its application.^[xii][12]  When we confront a new situation, we have to extend or modify our inheritance to encompass it, and as we do so, we thereby develop this inheritance.  Every time we attempt to apply a tradition, we have to reflect on it, we have to try to understand it afresh in the light of the relevant circumstances, and by reflecting on it, we necessarily open it up to possible innovation.  In this way, human agency can produce change even when people think they are adhering to a tradition they regard as sacrosanct.  As humans we necessarily reach our beliefs and perform our actions against a social background that influences those beliefs and actions.  As humans, however, we also possess a capacity for agency such that we can reason and act innovatively  against the background of tradition.

            It is the human capacity for agency that, I believe, makes tradition a more satisfactory concept than rivals such as structure, paradigm, and episteme.  All of these later concepts suggest the presence of a social force that determines, or at least sets  limits to, the performances of individuals.  The concept of tradition, in contrast, suggests that a social inheritance comes to each individual who, through their agency, then can modify and transform this inheritance even as they pass it on to yet others.  Although tradition thus seems to me to be preferable as a concept to others such as structure, we should not fetishize a particular word.  If other philosophers want to talk of decentered or open structures that allow for agency, then they will be discussing the same thing, only using different words.  What matters is that we have a proper conception of the relationship of the individual to society, not that we use a particular word to convey this conception.

Tradition and Ontology

            The concept of tradition captures an ontological fact or argument, that is, humans necessarily have their being in a social context which influences them.  How should we fill-out this ontological concept of tradition?  Because people necessarily arrive at their initial beliefs and practices by way of an inherited tradition, we can start by defining a tradition as a set of understandings someone acquires during a process of socialisation.  The new born infant develops into a mature adult with beliefs; and because these beliefs can not come exclusively from its experience and its reason alone, they must embody a tradition transmitted to it during a process of socialisation.  The new born infant learns to find its way in the world by being taught to recognise objects, identify their characteristics, name them, and speak about them, where the objects it thus picks out and discusses are made available to it by a tradition.  The infant learns to pick out objects as a result of being shown them and told their names, but what he or she can be shown and taught to name depends on the objects of which his or her teachers have experiences, and so on the theories with which these teachers already make sense of the world.  A tradition provides the theories that construct the objects the infant initially finds in the world.

            It is important to recognise that this analysis of the process of socialisation stands as a philosophical deduction from the grammar of our concepts, not an empirical induction from our theory-laden experience.  The empirical claim could be only that as a contingent fact people are embedded in social contexts.  The philosophical claim, in contrast, is that we can not conceive of anyone ever holding beliefs, and so performing actions, apart from in a social context.  Thus, there can not have been a time when traditions did not operate in this way: Leo Strauss must be wrong when he says that "classical philosophy originally acquired the fundamental concepts of political philosophy by starting from political phenomena as they present themselves to 'the natural consciousness.'"^[xiii][13]  Similarly, there could not be supermen capable of transcending the influence of tradition: Strauss also must be wrong when he implies that we might overcome tradition so as to perceive political phenomena in their natural appearance.^[xiv][14]  Nobody conceivably could escape the hold of tradition, neither someone from the past nor someone in the future.  Everyone at all times sets out from an inherited set of shared understandings that they acquire during a process of socialisation.  

            Although tradition is unavoidable, it is so as a starting point, not as something that determines, nor even limits, later performances.  We should be cautious, therefore, of representing tradition as an inevitable presence within all the individual ever does in case we thereby leave too slight a role for agency.  In particular, we should not imply that tradition is in some way constitutive of the beliefs that people later come to hold and the actions they then perform.  Although individuals must set out from within a tradition, they later can extend, modify, or even reject it in a way that might make it anything but constitutive of their later beliefs and actions.  Philosophers indebted to the hermeneutic tradition are particularly prone to talk of tradition, a social language, or something else of the sort, as if it were integral to everything the individual ever does.  They represent tradition as an impersonal force immanent within performances.^[xv][15]  Really, however, we should conceive of tradition primarily as an initial influence on people.  The content of the tradition will appear in their later performances only in so far as their agency has not led them to change it, where every part of it is in principle open to change.  Moreover, when the content of the tradition does not appear in their later performances, it influences them only by virtue of being the initial background against which they set out.  Tradition is an influence that works through individuals on individuals.  It is a necessary part of the background to everything anyone believes or does.  But it is not a necessary presence in all that they believe and do.

            Tradition is unavoidable as a starting point, not as a final destination.  This means that we need not define a tradition in essentialist terms - indeed we should be extremely wary of ever doing so.  The essentialist fallacy appears in A. O. Lovejoy's project of studying unit ideas as they change their outer form and enter into various shifting relationships with one another over time.  Lovejoy did not describe his unit ideas as traditions, but seems rather to have thought of unit ideas as appearing within traditions composed of clusters of unit ideas.  Nonetheless, whether we choose to talk of traditions, clusters of unit ideas, or unit ideas does not matter.  What matters is that we should eschew essentialism: we should be wary of any talk of "primary and persistent" objects.^[xvi][16]  Essentialists equate traditions with fixed essences to which they ascribe variations.  They define traditions in terms of an unchanging core - that appears in different outer garbs from time to time and even from person to person.  No doubt there are occasions when historians legitimately can point to the persistence through time of a core idea, or, rather, because, as we will see, the clarity of a tradition depends on the links between its constituent ideas, they can point to the persistence through time of a cluster of core ideas.  Equally, however, historians can choose to concentrate on a tradition with no essential core.  They might identify a tradition with a group of ideas widely shared by a number of individuals although no one idea was held by all of them.  Or they might identify a tradition with a group of ideas that was passed down from generation to generation, changing a little each time, so that no single idea persists from start to finish.

            Historians usually will encounter difficulties if they try to define a tradition in terms of a fixed core.  They will do so both because individuals are agents who play an active role in the learning process and because we can not identify limits to the changes individuals can introduce to their inheritance.  Individuals who take a given idea from their teachers do not also have to take the other ideas their teachers associate with it.  Rather, they can modify or reject any particular idea in any group of ideas whilst holding steady any of the other ideas in that group.  The bearer of a tradition might think of it as a unified whole possessing an essential core.  In fact, however, it will be composed of a variety of parts, each of which can be reflected upon, and so accepted, modified, or rejected, by itself.  Individuals can respond selectively to the different parts of the tradition they acquire as an inheritance.  Indeed, because people usually want to improve their heritage by making it more coherent, more accurate, and more relevant to contemporary issues, they often do respond selectively to it; they accept some parts of it, modify others, and reject others.  Traditions change as they are transmitted from person to person.^[xvii][17]  Essentialism typically results, therefore, either in a tradition whose range is severely restricted because agents rapidly depart from its fixed core, or in a tradition composed of beliefs that are defined so broadly that it lacks both clarity and explanatory power.

            Tradition is an influence that works through others on people, rather than a defining presence in all people believe and do.  The relationship of teacher to pupil provides a useful metaphor for the way in which others impart a tradition to someone, although we must be careful not to take this metaphor to refer exclusively to a formal, face to face relationship.  Individuals pick up their initial beliefs and practices by listening to and watching other people, including their parents, educators, the authors they read, and their peers. The learning process requires teachers who initiate and pupils who learn Typically each individual will fulfil both of these roles at some point in time.  The teachers once will have been pupils who acquired their initial beliefs and practices from earlier teachers, and the pupils later will become teachers who provide future pupils with initial beliefs and practices.  It is because beliefs and practices thus pass from generation to generation that we can talk of teachers initiating pupils into a tradition that persists and develops through time.  Although pupils receive their inheritance from teachers during fairly brief moments in time, these moments always represent the culmination of a larger historical process.  The teacher who transmits the inheritance is just the most recent link in a long chain of people who began as pupils and ended as teachers, passing an always changing set of beliefs and practices down to each other.  A long historical sequence lies behind the comparatively brief moment when a new pupil is initiated into a tradition.

            Because traditions persist only through teachers initiating pupils into shared understandings, we must avoid hypostatising them.  We must not ascribe to traditions an occult or Platonic existence independent of the beliefs and actions of specific individuals.  Traditions are not fixed entities people discover as already given.  They are contingent entities people produce by their own activities.  The exponents of a tradition bring it into being and determine its progress by developing their beliefs and practices in the ways they do.  Consequently, historians can identify the beliefs that make up a tradition only by reference to the shared understandings and temporal links that allow us to associate its exponents with one another.  Pupils learn what they do from individual teachers, not a social tradition: they listen to lectures by individuals, not society; they discuss affairs with individuals, not society; they read books written by individuals, not society; they watch television programmes made by individuals, not society; and they reflect on beliefs held by individuals, not society.  Intellectual traditions exist only as the sum of the beliefs of their individual exponents in their relations with one another.

            Let us examine more closely the nature of the relationship that must exist between beliefs and practices if they are to constitute a tradition.  For a start, the beliefs and practices that make up a tradition must have passed from generation to generation: they must embody a series of temporal relationships such that they provided the starting point for each of their later exemplars.  Traditions must be composed of beliefs and practices that were relayed from teacher to pupil to pupils' pupil and so on.  The existence of the appropriate temporal connections, however, need not have been a result of any deliberate design.  Nobody need have intended to pass on the relevant set of beliefs and actions, nor even have been conscious of doing so.  Typically the temporal continuity between the beliefs and practices within a tradition appears as a series of developments of transmitted themes.  As beliefs pass from teacher to pupil, so the pupil modifies and extends the themes, or conceptual connections, that linked the beliefs to each other.  Thus, although we must be able to trace a historical line from the start of a tradition to its current finish, the developments introduced by its successive adherents might be such that the start and finish have nothing in common apart from this temporal proximity.  The beliefs and actions of the most recent exponent of a tradition might be utterly different from those of earlier exponents of the same tradition.  However, an abstract set of beliefs and practices that was not passed on in the appropriate way would be a mere snapshot: it would be a summary of one or more moments in time, rather than a tradition capable of relating moments in time to one another by exhibiting their historical continuity.  If, for example, historians discovered that Chinese Buddhists and American Indians had held beliefs that resembled those of modern anarchists, they could not talk legitimately of a tradition of anarchism incorporating all these beliefs.  A tradition must consist of more than a series of instances that happen to resemble each other, or that resemble each other because they arose in similar situations or for similar reasons.  A tradition must consist of a series of instances that resemble one another precisely because they exercised a formative influence on one another in a definite temporal chain.

            As well as suitable temporal connections, traditions must embody suitable conceptual ones.  The beliefs and practices a teacher passes on to a pupil must form a fairly coherent set.  They must form an intelligible whole so that we can see why they went along together.  The fact is that the beliefs and actions of any individual must exhibit a minimal level of consistency, so the beliefs and practices within a tradition could not have provided someone with an initial starting point unless they coalesced to form a moderately coherent set.^[xviii][18]  Similar reasoning implies that the inner consistency of the beliefs and practices in a tradition must appear in their substantive content even if it also does so in a number of their other features, including an approach to certain objects, a mode of presentation, or an expression of allegiance.  Only beliefs and practices whose content cohered to some degree could provide the infant with an initial entry into the world.  Although the beliefs and practices in a tradition must thus exhibit a degree of conceptual coherence, this coherence need not be absolute.  Traditions can not be made up of purely random beliefs and actions that successive individuals happen to have held in common.  If, for example, historians discovered that various people believed both that God came to earth and that our souls survived death, they could not talk of a tradition composed of these beliefs alone.  If, however, they found that these two beliefs went along with others such as that Christ, the Son of God, came to earth and taught his followers to have faith in an afterlife, then they could talk of a Christian tradition composed of this fairly coherent set of beliefs.

            Although the beliefs within a tradition must be related to one another both temporally and conceptually, their substantive content is unimportant.  Because tradition is unavoidable, all beliefs and practices must have their roots in tradition.  They must do so whether they are aesthetic or practical, sacred or secular, legendary or factual, pre-modern or scientific, valued because of their lineage or their reasonableness.  It does not matter whether they are transmitted orally or in a written form.  It does not matter whether pupils are meant to accept them on another's authority or through being shown how to derive them from first principles.  All beliefs and all practices must arise against the background of tradition.  The ontological concept of tradition differs, therefore, from the one that some scholars associate exclusively with the customary ways of pre-modern peoples.  The concept of tradition we are analysing does not refer specifically to pre-industrial, rural communities governed by prescriptive authority and customary laws inspired by religious values.  It applies equally cogently to modern communities with their greater concern for legal authority and rational laws ostensibly grounded on scientific knowledge.  Perhaps there is a useful distinction to be made between pre-modern and scientific authorities, between beliefs adopted as part of an entrenched folklore and those adopted as a result of methodical procedures and controlled reasoning.  Even if there is, however, it must occur within the ontological concept of tradition.^[xix][19]  Because even we moderns can not have pure experiences, we too must arrive at our beliefs and actions by way of the traditions found in our communities.  Novices in modern science do not work out appropriate procedures, reasoning, and accepted truths by themselves.^[xx][20]  Rather they are initiated into a tradition of science by their teachers, and only after they have been thus initiated do they proceed to advance science through their own work.  Even when they later go back to check or repudiate received scientific wisdom, they do so against the background of a tradition into which they already have been initiated.

Traditions and Explanation

            The ontological concept of tradition captures a very general fact about the human condition.  It belongs to the categories or presuppositions of historical scholarship in that it tells us about the nature of the concepts and the forms of explanation that are appropriate to the study of the past.^[xxi][21]  As such, however, the ontological concept of tradition provides little immediate help in the process of constructing particular historical theories or explanations.  What procedures of idealization should we use to construct a tradition to explain a particular belief or action?  Historians have before them, first, individuals who hold beliefs and perform actions, and, second, traditions composed of beliefs and practices that are related to each other in a suitable temporal and conceptual manner.  How should historians identify, or construct, a particular tradition to explain why a particular individual holds certain beliefs and performs certain acts?  Many of the problems attendant on the concept of tradition arise because historians try to answer this question by comparing the beliefs and actions of the individual with a reified tradition.  Just as we rejected an essentialist analysis of tradition, so we must eschew the temptation to locate individuals in a tradition by comparing their beliefs and actions with a checklist of core ideas, a suitable philosophical move, or any other allegedly defining feature of a tradition.  Because traditions are not fixed entities of which specific instances partake, we can not locate people in one by comparing their beliefs and actions with its allegedly key features.  Rather, because traditions are the contingent products of the ways in which people develop specific beliefs and practices, we must identify the tradition against the background of which people come to hold the beliefs they do by tracing the appropriate temporal connections back through time.

            We should not reify traditions, as if the question to ask were "does this individual express the ideas or engage in the practices constitutive of this fixed tradition?" rather than "what temporal connections do we need to trace back to explain why this individual had these beliefs and practices as a starting point?"^[xxii][22]  Similarly, we should not regard a tradition as "a concrete manner of behaviour" as if there were some authentic set of experiences that its exponents attempt to articulate as constitutive of it.^[xxiii][23]  Traditions are not fixed entities that play a judicial role in our understanding: they do not enable us to evaluate particular beliefs and actions against an allegedly privileged set of beliefs or an allegedly authentic set of experiences and actions.  Rather, traditions are evolving entities that play an instrumental role in our understanding: they help us to explain a particular belief or action by relating it to relevant prior beliefs or actions.

            No particular set of beliefs, experiences, or actions has a privileged, automatic, or natural role in defining any tradition.  Even if Marx himself somehow told us he had meant such and such, this would not mean that his expression constituted a privileged, defining set of beliefs for a hypostatised Marxist tradition.  The Marxist tradition in which we locate someone must consist of whatever beliefs we come across as we trace the temporal connections back from that person's beliefs in our attempt to explain them.  Marx's new statement would be just one more that we might or might not arrive at whilst tracing these connections.  There is no single, authentic Marxist tradition, just numerous Marxist traditions, each of which helps to explain a different person's beliefs.  Likewise, no particular set of logical relationships has a privileged, automatic, or natural role in defining any tradition.  Historians must define traditions in terms of beliefs that were related to one another in an appropriate manner.  They can not do so in terms of the logical relationships they believe relate various beliefs to one another.  Even if Locke held beliefs that we can read as responses to, or elucidations of, Hobbes's beliefs, Hobbes still would not necessarily belong in the tradition in which we locate Locke.  The tradition in which we locate Locke must consist of the beliefs that we come across as we trace temporal connections back from his beliefs in our attempt to explain them.  The logical links between Locke and Hobbes become relevant only if they are reinforced by appropriate temporal one.  Historians can identify a tradition only through a study of the beliefs and actions of the individuals within it.  Only the beliefs and actions of individuals can acquaint them with traditions; only inference from the beliefs and actions of individuals can enable them to explore the nature of traditions; and only checks against the beliefs and actions of individuals can provide them with tests of their claims about traditions.

            Because traditions are not hypostatised entities, a historian can decide which individuals belong in a tradition only by tracing the temporal connections that bind a particular belief or act back to its predecessors.  A historian can not decide whether individuals belong in a tradition by comparing their beliefs and actions with an abstract moment in a logical argument or with a privileged set of beliefs, experiences, or actions.  Even if people want to identify themselves with a tradition, they can not do so by saying that their beliefs and actions share key features with, or address questions raised by, those they see as their predecessors.  People can identify themselves with a tradition only by showing that they are linked to those they see as their predecessors by a series of appropriate temporal connections.  Whenever people locate themselves in a tradition, therefore, they make a historical argument with which others might disagree.  To locate themselves in a tradition, they have to defend a particular account of the conceptual and temporal relationships between the beliefs and practices of those they see as their predecessors.  Moreover, this account usually will conflict with the accounts other people give of these relationships when locating themselves in a similar or different tradition.  Hence conflicts over how to interpret traditions are a more or less permanent feature of social life.  Such conflicts arise because people offer incompatible views of the past in their attempts to locate themselves in traditions.  Such conflicts do not imply that there is an authentic tradition over which to struggle.  Nor do they, together with the developments they produce, constitute the defining characteristic of traditions.^[xxiv][24]

            A rejection of all hypostatised views of tradition should lead us to conclude that historians can locate an individual in a variety of traditions depending on their different purposes.  Because there are no hypostatised traditions, the historian's task can not be to locate the individual in one of a finite set of fixed traditions.  Rather, because historians identify the tradition against which someone believed or did something by tracing the relevant temporal connections back through time, the precise content they give to the tradition will depend on the particular beliefs or actions they thereby hope to explain.  If they want to explain someone's beliefs and actions as a whole, they will define the relevant tradition in one way; but if they want to explain only part of these beliefs and some of these actions, they might define the relevant tradition differently; and if they want to explain another part of these beliefs, they might define the relevant tradition differently again.  Because historians identify a tradition by tracing temporal and conceptual links back through time starting with the particular instances they wish to explain, the content of the tradition they identify typically varies with the instances they wish to explain.  Different features of a person's beliefs and actions typically have somewhat different temporal and conceptual connections to those of other people.  Thus, when two historians set out to explain different features of a person's life or work, they typically trace back different temporal and conceptual connections in a way that quite properly leads them to identify slightly different traditions as appropriate explanatory contexts for the different objects of interest to them.  There is a very real sense, therefore, in which historians define traditions according to their own purposes.  Historians do not pigeon-hole each individual in one of a fixed number of reified traditions.  They select one tradition from among the many in which they could locate an individual because it best explains the particular beliefs and actions they are studying.

            There is a sense in which historians construct traditions for themselves by picking out the beliefs and habits of appropriate individuals using criteria of relevance deriving from their own interests.  But this need not worry us.  Any abstraction we make will depend on a principle of classification the rationale for which derives from our purpose in making it.  The fact that we construct traditions does not imply that they are unacceptably subjective.  Their objective nature depends on the adequacy of our understanding of the beliefs and practices we classify as part of them, not on the principle by which we classify these beliefs and practices.  An account of a tradition must identify a set of connected beliefs and habits that intentionally or unintentionally passed from generation to generation at some time in the past.  Although historians can construct a tradition to suit the purpose of their inquiry, the tradition they construct must have existed, so they must show that individuals really did hold the beliefs and habits of which it is composed.  Moreover, if historians want to demonstrate that someone was self-consciously a part of a tradition, they also must show that this person defined the tradition more or less as they do.

            Once we recognise that historians can select traditions to suit their different purposes, we will dismiss as besides the point several heated debates about how exactly historians should make sense of particular individuals.  Consider, for example, the complaints made by contextualists about the failure of epic theorists to locate authors in a proper historical context.  Consider, more particularly, John Dunn's complaint about Strauss's failure to locate Locke in the Puritan tradition.  We have found that what counts as a proper historical context depends on what one hopes to explain.  This suggests that the interpretations of the epic theorists and their critics often are more compatible than either group appears to think.  For example, if, like Dunn, historians want to explain Locke's political thought, then no doubt Locke's debt to the Puritan tradition will be of much greater import than his debt to Hobbes; but if, like Strauss, historians want to explain features of modern political philosophy found in Hobbes and Locke, then no doubt they should construct a tradition rather different from the Puritan one evoked by Dunn.^[xxv][25]  Even if the tradition Strauss describes does not provide the best context for Locke's thought as a whole, it still might be the right one to explain the features of Locke's thought in which he is interested.  Dunn's complaint that Strauss ignores Locke's debt to the religious ideas of the Puritans misses the point.  Properly to repudiate the traditions constructed by epic theorists such as Strauss, contextualists must show that these traditions do not embody appropriate temporal and conceptual connections.  They should condemn Strauss not by arguing that he misinterprets certain authors, but, as J. G. A. Pocock, Quentin Skinner, and others have done, by arguing that the tradition he postulates lacks suitable temporal connections.^[xxvi][26]  All too often epic theorists evoke an alleged tradition of classic works that run from Plato onwards without bothering to demonstrate the historical existence of genuine temporal links between the works they include in this tradition.

            Once we recognise that historians can select traditions to suit their different purposes, we also will deny that historians always must define a given epoch in terms of a single tradition, episteme, or whatever.  To reject the hypostatisation of traditions is to imply that we can not say of traditions what Michel Foucault does of epistemes, that is, that "in any given culture and at any given moment, there is always only one episteme."^[xxvii][27]  Earlier we found that traditions do not define the instances within them since pupils are agents who play an active part in the learning process.  Now we have found that historians select traditions to fit the particular instances they wish to explain.  Clearly, therefore, there is no reason why historians should have to select traditions that cover the whole of an epoch.  Individuals disagree as well as agree, so historians always can pick out a plurality of traditions that were present at any given time.  Moreover, because individuals disagree at various levels of generality, historians can choose how broadly they want to define their traditions.  No doubt historians can pick out very general themes characteristic of the whole of an epoch, and, moreover, they then might describe the result as something such as an episteme, though not an episteme that precludes agency.  Equally, however, historians can pick out themes that are found only in this group, or only in that group, and, moreover, they then can describe the result as a number of overlapping, competing traditions.  We can conclude, therefore, that the thought and practices of an epoch can not possesses a monolithic character that precludes historians talking of there being more than one tradition at work therein.

            If historians do identify a single tradition or episteme governing the whole of an epoch, that tradition will be of little interest since it will have little explanatory power.  An epoch is made up of the beliefs and actions of numerous individuals complete with all their agreements and disagreements.  Historians select a tradition from within this medley of belief and action in order to explain a particular set of instances.  The explanatory value of traditions lies in the way they illustrate the process by which individuals inherited beliefs and practices from their communities.  Thus, the wider historians define a tradition, the weaker its explanatory power will be.  If historians select monolithic epistemes, they will have to define them solely in terms of the beliefs held and the actions performed by everyone in an epoch, so when they try to explain the beliefs and actions of particular individuals, they will be able to explain only why they held these universal beliefs and performed these universal actions, not why they held numerous other, more specific beliefs and why they performed numerous other, more specific action.

            The narrower the definition historians give of a tradition, the greater will be its explanatory power.  Historians pick out specific historical traditions from the general tradition in which we all have our being in order to show how the process of inheritance worked with respect to specific instances.  They select traditions out of tradition, conceived as the general background of human life, in order to explain specific features of that life.  The value of the traditions they select derives from the explanatory power of the conceptual and temporal links found between the beliefs and actions of which the traditions are composed.  The clarity and precision with which historians analyse these links fixes the intelligibility and relevance of the tradition they select.  The more exact their account of these links, the more fully we will be able to grasp the nature and location of the tradition, so the more explanatory work it will be able to do.  Temporal links reveal a movement from the beliefs of teachers to those of pupils.  They show how the relevant beliefs and practices passed from one generation to another thereby explaining why the beliefs persisted through time.  Conceptual links reveal a pattern in beliefs that persisted together through time.  They show us how the relevant beliefs and practices form a fairly coherent set thereby explaining why they persisted together as a loose knit whole rather than as isolated units or units brought together by mere chance.

            Our discussion of tradition and explanation has been conducted at a general level.  We have scarcely approached an answer to specific issues such as "Is Marxism a tradition?" or "was Jeremy Bentham the teacher of J. S. Mill?".  Nor have we provided clear criteria by which to decide such issues.  The worry here is that an avoidance of specific cases points to an evasion of questions such as "how much variation is compatible with the presence of a tradition?" and what evidence do we need to establish the presence of a teacher-pupil relationship?"  How should we answer these questions?  The first thing to note is the location of minimal criteria within our account of tradition.  The beliefs within a tradition must form an intelligible whole that we can recognise as such, as, on most accounts, Marxism does.  Similarly, a pupil must share ideas with a teacher from whom he could have acquired them, as, on most accounts, Mill does with Bentham.  The second thing to note is the insufficiency of these minimal criteria.  Because we construct traditions to explain a particular later instance, the question is not simply whether or not the relevant links can be shown to exist, but also whether or not they are the ones that provide the best explanation of the later instance.  We need to show that the tradition we postulate is the most helpful one for an explanation of the beliefs or habits we want to explain.  The final thing to note is that despite the insufficiency of our minimal criteria we can not specify stronger ones.  Because the value of any tradition we construct depends on its being the most helpful one for an explanation of a particular instance, whether or not we are justified in postulating it must depend on a comparison with the available alternatives, not just an evaluation against independent criteria.  We can not decide specific issues such as "is Marxism a tradition?" and "was Bentham the teacher of Mill?" solely by reference  to the evidence and a theory of tradition.  However, to leave the specifics open in this way is not to evade them precisely because our theory of tradition establishes that we have to leave them open in this way.

            Tradition constitutes the inescapable background to human life.  Historians construct particular traditions out of the general flux of tradition by tracing the temporal and conceptual connections that flow out of the particular object or objects that they want to explain.  What are the nature and limits of the role of traditions in explaining particular instances?  Earlier we found that although tradition is the unavoidable background to all we say and do, but not a constitutive presence in all we say and do.  Traditions are not hypostatised entities which appear in various guises at different times.  They are, rather, contingent and evolving entities that operate through teachers as influences on pupils, where the pupils then can extend and modify them in unlimited ways.  The role of traditions, therefore, must be to explain why people set out with the beliefs and practices they did, not to explain why they went on to change these initial beliefs and practices in the ways they did.  For a start, because pupils sometimes remain faithful to their inheritance, they sometimes hold to beliefs and practices that correspond to a tradition imparted to them by others.  Whenever pupils learn something from a teacher, one way of explaining the beliefs and actions of the pupils is to say that they learnt them from a teacher.  Thus, historians sometimes can explain why people believe or do something simply by saying that they learnt it from teachers who imparted a tradition to them.  More importantly, because no belief or action can be self-supporting, individuals always must locate their particular beliefs and actions in a larger set.  Pupils must acquire a set of beliefs and actions in an initial process of socialisation before they can modify this set.  Whenever pupils depart from their inheritance, they necessarily do so against a background composed of the tradition their teachers imparted to them.  Thus, historians who want to explain the development of people's beliefs and practices must set out from the tradition from which their subjects began.

            When historians identify a tradition as a starting point for an individual, they describe that person's beliefs and actions in relation to various other beliefs and actions that they have selected from among the complex totality of the past.  They associate the imagery, syntax, phraseology, and conceptual content of that person's beliefs and actions with those of an identifiable set of prior beliefs and actions.  When they do so, they show that person to have subscribed to a tradition composed of this set of prior beliefs and actions, where, of course, to subscribe to a tradition is not to follow it slavishly but rather to set out from it.  Moreover, when historians show a tradition was the point of departure for an individual, they identify it as a suitable point of departure for their explanation of why that person's later development.  To say that traditions provide historians with points of departure from which to explain something is, however, to recognise that traditions are not sufficient for such explanations.  Because traditions evolve as a result of human agency, they can not explain the ways in which people develop the relevant beliefs and practices.  Historians can not invoke traditions to explain why people modified their inheritance in the way they did; they can not do so because the modification of an inheritance is an act of agency performed against the background of a tradition but not decided by it.  Fully to explain why people believed and did what they did, therefore, historians must supplement explanations that refer to traditions with explanations that refer to agency.

Tradition and Ethics

            Tradition is the ineluctable, albeit diffuse, background to human beliefs and actions.  As such, tradition offers the historian a vehicle of explanation: the historian can explain a belief, action, set of beliefs, or practice at least in part by locating it within the tradition informing it.  Despite the usefulness of the concept of tradition, however, many historians remain suspicious of it.  Perhaps the most important source of their suspicions is the political uses to which conservatives put the concept.  To conclude, therefore, I want briefly to explore the ethical or political implications of the concept of tradition.  The key point to make here is that the ineluctable nature of tradition precludes it being of itself a good or bad thing.  Given that we can not escape the influence of tradition, the idea that we should strive to preserve it surely must seem an odd one.  The ineluctable nature of tradition suggests, rather, that the proper ethical question to ask is: what sort of tradition should we promote?  Conservatives typically advocate comparatively closed and hierarchic traditions.  They defend traditions that are hostile to extensive innovation and that evoke an authoritative elite as a bulwark against such innovation.  Indeed, the zeal with which conservatives promote such traditions often leads them to invent authorities, symbols, and pageants so as to buttress what is in fact a relatively new tradition.  Radicals, however, legitimately might advocate comparatively open and egalitarian traditions quite different from those favoured by conservatives.  Knowing that humans are capable of agency, for example, radicals might promote traditions that explicitly recognise this capacity and even encourage their adherents to innovate.^[xxviii][28]  Conservatives, of course, often suggest that excessive innovation is unacceptably disruptive of social order.  Clearly, however, whether or not this suggestion is right remains a matter of sociological judgement; after all, radicals reasonably might suggest that too restrictive an environment is at least as disruptive of social order as is excessive innovation.

            To acknowledge the ontological fact, and explanatory role, of tradition is by no means to commit oneself to a conservative ethics or politics.  Historians should recognise that individuals come to hold beliefs and to make actions only against the background of tradition, where traditions influence but do not determine or limit these beliefs and actions.  Historians should construct appropriate traditions to explain the ideas, events, and practices of the past.