In other words , he has no adequate basis on which to determine whether , in consequence of the deviance he refers to , there was , or was not , a stronger definition of the moral and social boundaries of the community .sx So far as popular perceptions and evaluations are concerned , he is without means of access .sx Likewise , in treating the second hypothesis , on the constant level of deviance , Erikson has to rely on official crime statistics , which , for well-known reasons , give only a very uncertain indication of the actual level of social deviance , and are influenced in their trend by a variety of other factors .sx However , unlike the sociologist of deviance working in contemporary society , Erikson cannot investigate in any detail the processes through which the official statistics were constituted , nor can he collect data of his own which could provide alternative estimates - as , say , through some form of 'victim survey' .sx To be sure , the hypotheses that Erikson addresses are not ones that would be easily tested under any circumstances .sx But , given that they derive from a theory that pretends to a very high level of generality , there is all the more reason to ask why Erikson should impose upon himself the limitations that must follow from choosing a historical case .sx Why should he deny himself the possibility of being able to generate his own evidence , to his own design , and under conditions in which problems of reliability and validity could best be grappled with ?sx Any sociologist , I would maintain , who is concerned with a theory that can be tested in the present should so test it , in the first place ; for it is , in all probability , in this way that it can be tested most rigorously .sx I would now like to move on to consider cases where the recourse of sociologists to history would appear to have the good reasons which , I earlier maintained , should always be present .sx Here my aim is to illustrate what such reasons might be , but also - when they are acted upon - the difficulties that may be expected .sx Sociologists , one might think , will most obviously need to turn to history where their interests lie in social change .sx However , it should be kept in mind that a recourse to the past - or , that is , to the relics thereof - is not the only means through which such interests may be pursued :sx life-course , cohort or panel studies , for example , are all ways of studying social change on the basis of evidence that is , or has been , collected in the present .sx Sociologists , I would argue , are compelled into historical research only where their concern is with social change that is in fact historically defined :sx that is , with change not over some analytically specified length of time - such as , say , 'the life-cycle' or 'two generations' - but with change over a period of past time that has dates ( even if not very precise ones ) and that is related to a particular place .sx Sociologists have a legitimate , and necessary , concern with such historically defined social change because , as I have earlier suggested , they wish to know how widely over time and space their theories and hypotheses might apply .sx One illustration of what I have in mind here is provided by Michael Anderson's book , Family Structure in Nineteenth Century Lancashire .sx Anderson is concerned with the hypothesis that in the process of industrialisation , pre-existing forms of 'extended' family and kinship relations are disrupted .sx Specifically , he is interested in whether or not this hypothesis holds good in the British case - that of the 'first industrial nation' .sx Thus , to pursue this issue , Anderson aims to examine just what was happening to kinship relations in Britain at the time when , and in the place where , the 'take-off' into industrialism is classically located .sx In contrast , then , with Erikson , Anderson has a quite clear rationale for turning to historical research .sx A second illustration is provided by Gordon Marshall's book , Presbyteries and Profits .sx Marshall is concerned with the 'Weber thesis' - that a connection exists between the secular ethic of ascetic Protestantism and 'the spirit of capitalism' .sx In the long-standing debate on this thesis , the case of Scotland has several times been suggested as a critical one , in that , in the early modern period , Scotland had a great deal of ascetic Protestantism - that is , Calvinism - yet showed little in the way of capitalist development .sx Marshall's aim is then to re-examine the Scottish case for the period from around 1560 down to the Act of Union of 1707 .sx Marshall points out that Weber himself always emphasised that his argument on the role of the Protestant ethic in the emergence of modern capitalism was intended to apply only to the early stages of this process :sx once a predominantly capitalist economy was established , its own exigencies - in the workplace and market - would themselves compel behaviour generally consistent with the 'spirit of capitalism' without need of help from religion .sx Again , then , Marshall , like Anderson , has obviously good grounds for his recourse to history .sx Now before proceeding further , I should make it clear that I have the highest regard for the two studies to which I have just referred .sx Both make signal contributions to the questions they address ; and , for me , they stand as leading examples of how in fact historical sociology should be conceived and conducted .sx I say this because I want now to go on to emphasise the severe limitations to which the analyses of both authors are subject :sx not because of their deficiencies as sociologists , but simply because of the fact that they were forced into using historical evidence - forced into a reliance on relics - rather than being able to generate their own evidence within a contemporary society .sx The relics on which Anderson chiefly relies are the original enumerator's books for the censuses of 1841 , 1851 and 1861 .sx On this basis , he can reconstruct household composition according to age , sex and kinship relations , and he can also to some extent examine the residential propinquity of kin .sx But this still leaves him a long way short of adequate evidence on the part actually played by kinship in the lives of the people he is studying and on the meanings of kinship for them .sx He attempts to fill out the essentially demographic data that he has from the enumerators' books by material from contemporary accounts .sx But these would , I fear , have at best to be categorised as 'casual empiricism' and at worst as local gossip or travellers' tales .sx Titles such as Walks in South Lancashire and on its Borders , A Visit to Lancashire in December 1862 , and Lancashire Sketches give the flavour .sx Anderson is in fact entirely frank about the problem he faces .sx " It must of course be stressed " , he writes , " that just because interaction with kin occurred it is no necessary indication that kinship was important .sx The real test , which is quite impossible in any precise way in historical work , would be to examine the extent to which kinship was given preference over other relational contacts ( and the reasons for this preference ) , and the extent to which contacts with kin fulfilled functions which were not adequately met if kin did not provide them " .sx The point I want to make here would perhaps best be brought out if one were to compare Anderson's study of kinship with one carried out in contemporary society - let us say , for example , Claud Fischer's study of kinship and of other 'primary' relations in present-day San Francisco , To Dwell Among Friends .sx The only conclusion could be that the latter is greatly superior in the range and quality of data on which it draws , and in turn the rigour and refinement of the analyses it can offer .sx And this point is , of course , not that Fischer is a better sociologist than Anderson but that he has an enormous advantage over Anderson in being able to generate his own data rather than having to rely on whatever relics might happen to be extant .sx Turning to Marshall , one finds that he has problems essentially the same as those of Anderson .sx One of Marshall's main concerns is that Weber's position should be correctly understood - following the vulgarisations of Robertson , Tawney , Samuelson and other critics ; and in this respect Marshall makes two main points .sx First , Weber was not so much concerned with official Calvinist doctrine on economic activity as with the consequences of being a believing Calvinist for the individual's conduct of everyday life - consequences which the individual might not even fully realise .sx In other words , Weber's thesis was ultimately not about theology but subculture and psychology .sx Secondly , Weber's argument was that the Protestant ethic was a necessary , but not a sufficient cause of the emergence of modern capitalism ; there were necessary 'material' factors also - such as access to physical resources and to markets , the availability of capital and credit etc. Thus , Marshall argues , in evaluating the Weber thesis , it is not enough to look simply for some overt association between theology , on the one hand , and the development of capitalist enterprise on the other .sx What is required is more subtle .sx It is evidence that believing Calvinists , on account of their acceptance of a Calvinist world-view , were distinctively oriented to work in a regular , disciplined way , to pursue economic gain rationally , and to accumulate rather than to consume extravagantly - so that , if other conditions were also met , capitalist enterprise would then flourish .sx Marshall's position here is , I believe , entirely sound .sx But it leads him to problems of evidence that he can in fact never satisfactorily overcome - despite his diligence in searching out new sources and his ingenuity is using known ones .sx And the basic difficulty is that relics from which inferences can systematically be made about the orientations to work and to money of early modern Scots are very few and far between .sx In other words , what is crucially lacking - just as it was lacking for Anderson and indeed for Erikson - is material from which inferences might be made , with some assurance of representativeness , about the patterns of social action that are of interest within particular collectivities .sx As Clubb has observed , the data from which historians work only rarely allow access to the subjective orientations of actors en masse , and inferences made in this respect from actual behaviour tend always to be question-begging .sx And Marshall , it should be said , like Anderson , sees the difficulty clearly enough .sx He acknowledges that it may well be that " the kind of data required in order to establish the ethos in which seventeenth-century Scottish business enterprises were run simply does not exist " - or , at least , not in sufficient quantity to allow one to test empirically whether Calvinism did indeed have the effect on mundane conduct that Weber ascribed to it .sx III Let me at this point recapitulate .sx I have argued that history and sociology differ perhaps most consequentially in the nature of the evidence on which they rely , and that this difference has major implications for the use of history in sociology .sx I have presented a case of what , from this standpoint , must be seen as a perverse recourse to history on the part of a sociologist ; and I have now discussed two further cases where , in contrast , such a recourse was justifiable , indeed necessary , given the issues addressed , but where , none the less , serious difficulties arise because of the inadequacy of the relics as a basis for treating these issues .sx To end with , however , I would like to move on from these instances of sociologists resorting to history in the pursuit of quite specific problems to consider - with my initial argument still in mind - a whole genre of sociology which is in fact dependent upon history in its very conception .sx I refer here to a kind of historical sociology clearly different to that represented by the work of Anderson or Marshall , and which has two main distinguishing features .sx First , it resorts to history because it addresses very large themes , which typically involve the tracing out of long-term 'developmental' processes or patterns or the making of comparisons across a wide range of historical societies or even civilisations .sx