citação
Proposições de Law e Lodge (1984)

Proposições de Law e Lodge (1984)

Network theory axioms by Law & Lodge (1984)

Segue abaixo a transcrição das 161 proposições que John Law e Peter Lodge distribuíram ao longo dos 24 capítulos de Science for social scientists (Law e Lodge 1984). As proposições estão na sequência original, mas foram organizadas por mim em temas que não correspondem exatamente à divisão original dos capítulos do livro.

CLASSIFICATION
[1] [Classification] is something for which human beings have a natural psychological capacity. (Law e Lodge 1984:23)
[2] Part of this capacity is built into our brains. We notice and associate certain kinds of stimuli. (Law e Lodge 1984:23)
[3] We do this unconsciously, using stimuli which come to us from the outside world. (Law e Lodge 1984:23)
[4] Part of the capacity is learned. There are culturally prescribed classes into which we slot phenomena. (Law e Lodge 1984:23)
[5] This too is normally an unconscious phenomenon. (Law e Lodge 1984:23)
[6] It depends on the recognition of similarity between objects, though the nature of that similarity may be obscure. (Law e Lodge 1984:23)
[7] All classification involves loss of information. (Law e Lodge 1984:23)

ASSOCIATION (cross-classification)
[8]  Human beings are capable of associating classes together in a manner identical to their ability to notice similarity between the objects that go to make up a class. (Law e Lodge 1984:30)
[9]  Much of this cross-classification operates below the level of consciousness – again it is perceptual. (Law e Lodge 1984:30)
[10] Cross-classification can be learned either directly (via empirical association) or indirectly (via linkages with other already usable classes). (Law e Lodge 1984:30)

LEARNING ASSOCIATIONS
[11] Even indirect learning has in the end to be tied down to the empirical via chains of classificatory associations. (Law e Lodge 1984:30)
[12] All directly or indirectly learned associations may be either true or false. (Law e Lodge 1984:30)
[13] Cross-classifications may take the form of predictions. (Law e Lodge 1984:30)

NETWORKS
[14] Cross-classification leads to a densely interrelated set of classes which, following [Mary B.] Hesse, we call a ‘network’. (Law e Lodge 1984:43)
[15] Despite the complexity of networks these are ordered by a search for simple, economical and general classes and interconnections. (Law e Lodge 1984:43)
[16] This does not imply that there is a single ‘best’ classification for all individuals. (Law e Lodge 1984:43)

MISFITS and NETWORK ALTERATION
[17] In the search for economy of classification misfits occur. (Law e Lodge 1984:43)
[18] In the attempt to eradicate misfits, all terms and links in a network are at risk for alteration. (Law e Lodge 1984:43)
[19] This includes terms with a direct empirical referent, because such terms rest upon a small selection from an indefinite number of possible similarity relations. Other similarities might be chosen. (Law e Lodge 1984:43)
[20] It follows that empirical descriptions are not dependent on nature alone (though nature makes them possible). (Law e Lodge 1984:43)

MEANING (intension & extension)
[21] The meaning of any term depends, in addition, upon its position in a network, both with respect to the objects referred to by that term (extension) and the properties with which those objects are seen as being similar (intension). (Law e Lodge 1984:44)
[22] Agreement about extension does not entail agreement about intension. (Law e Lodge 1984:44)
[23] The extension and intension of a term are both susceptible to alteration as the result of changes elsewhere in the network. (Law e Lodge 1984:44)

COHERENCE CONDITIONS (economy, maximizing; cultural conventions)
[24] Association between classes can take a variety of forms: definition, law, empirical generalisation, empirical finding, indicator or tacit perceptual assumption. (Law e Lodge 1984:52-3)
[25] These forms are not rigidly distinct. One type of association may turn into another if the conditions are right. (Law e Lodge 1984:53)
[26] Coherence conditions are the principles by which a network is ordered and structured. (Law e Lodge 1984:53)
[27] The most basic coherence condition is economy. Human beings try to simplify their mental maps by detecting similarity between new experience and old classes. Thus, they try to maximise the ordering power of their existing networks. (Law e Lodge 1984:53)
[28] Naturally misfits may be perceived. Then network change (either to the old classes or the new phenomenon) may take place. (Law e Lodge 1984:53)
[29] The coherence condition of economy does not determine a unique network for all human beings because the latter experience different biographies, have different interests and have different explanatory principles made available to them in the course of socialisation. (Law e Lodge 1984:53)
[30] Such socially transmitted explanatory principles are also coherence conditions. They include preferred models and modes of explanation. (Law e Lodge 1984:53)
[31] It follows that all knowledge, including scientific knowledge, is constitutively conventional. (Law e Lodge 1984:53)

LINK PROBABILITY
[32] The links in a network vary in strength. This strength may for certain purposes be measured as the probability of that link. (Law e Lodge 1984:60)
[33] Confirmation theory has attempted to quantify that probability. In particular, personalist probability theory proposes that the strength of a link be treated as the rational degree of belief in that link. (Law e Lodge 1984:60)
[34] This may be calculated in a number of ways, though Bayes’s theorem has been particularly useful in attempting to mimic the acquisition of knowledge by a ‘human learning-machine’. (Law e Lodge 1984:60)
[35] Given two classes, A and B, the probability that A is associated with B may or may not be the same as the probability that B is associated with A. (Law e Lodge 1984:60)
[36] Probabilities may take any value between plus one (identity, invariant empirical link, complete correlation) and zero. (Law e Lodge 1984:61)
[37] Over time the links between classes may vary in terms of probability. (Law e Lodge 1984:61)
[38] All classes in a network may, in principle, be linked with all others. (Law e Lodge 1984:61)

LINK UTILITY
[39] However, a principle of economy is at work here. Links are not actually constructed unless they are of some use. (Law e Lodge 1984:61)
[40] The utility of a link between classes is partially independent of its probability value(s). (Law e Lodge 1984:61)

PRAGMATIC NETWORK ADEQUACY-ACCURACY
[41] The adequacy of a map or plan depends upon the purposes for which it is used. There is no such thing as a completely accurate map for all purposes. (Law e Lodge 1984:71)
[42] Even apparently accurate photographs vary in the extent to which they are workable from any given standpoint. (Law e Lodge 1984:71)
[43] The same is true for networks. They do not completely represent reality. Rather, they depict it for certain purposes. (Law e Lodge 1984:71)
[44] The reason for this is that the construction of networks rests upon the selection of a small number of similarities and associations from the vast range of possibilities. There is always a loss of information. (Law e Lodge 1984:71)
[45] The loss of information involved in network construction means that if the network is used for other purposes it is likely to prove inadequate. (Law e Lodge 1984:71)
[46] The network theory is thus in accord with instrumental or pragmatic rather than correspondence theories of knowledge. (Law e Lodge 1984:72)

DESCRIPTION & PRESCRIPTION (truth)
[47] Given that the term ‘truth’ is closely associated with correspondence theory, it might, on balance be better to abandon its use. (Law e Lodge 1984:72)
[48] Our account of the generation of knowledge, unlike that of most philosophies of science, is descriptive in intent. (Law e Lodge 1984:83)
[49] Nevertheless certain implications of the network model suggest that some features of prescriptive philosophies of science are unrealistic. (Law e Lodge 1984:83)

COHERENCE CONDITIONS (críticas a Mills e a Popper)
[50] Thus, the network theory suggests that all data are sensitive not only to perceptual ‘lumpiness’ but also to the operation of coherence conditions. (Law e Lodge 1984:83)
[51] This is incompatible with those philosophies of science – for instance, Mills’ inductivism and some versions of Popper’s falsificationism – which rest upon the contrary assumption that data can be constructed in the absence of coherence conditions. (Law e Lodge 1984:83)
[52] In practice a test of theory against data must be seen as a test of one kind of coherence condition (theory) against another (theoretically-selected perceptual lumpiness). (Law e Lodge 1984:83)
[53] The network model also suggests that it is impossible to abandon all knowledge simultaneously, though radical but piecemeal reconstruction of the whole is quite possible. This is incompatible with Popper’s insistence that falsified theory be rejected outright. (Law e Lodge 1984:83)
[54] Despite its failure to consider coherence conditions, Mill’s theory of induction is correct in one important respect – its stress on the importance of the detection of similarity and association. (Law e Lodge 1984:83)
[55] Likewise, Popper’s theory which stresses the role of theoretical activity (as opposed to induction) is also correct in this respect. (Law e Lodge 1984:84)
[56] Finally, network theory suggests that two-language conceptions of science which draw a basic distinction between terms and axioms, dictionaries and empirical observations, overstress such distinctions. In practice, though distinctions such as these do exist, those who are primarily concerned with the description of conceptual change are better advised to treat them as practical differences which may change over the course of time. (Law e Lodge 1984:84)

LEARNING & AUTHORITY
[57] Learning involves the discrimination of perceptual objects into classes. (Law e Lodge 1984:96)
[58] Authority provides cues about conventional discriminations (though authority in turn is built upon the basis of workable perceptual cues). (Law e Lodge 1984:96)
[59] Perceptual grouping (‘seeing’) is not something that is given in the perceptual lumpiness of the world alone. (Law e Lodge 1984:96)
[60] Once a usable perceptual classification has been established it is remarkably difficult to break this down. (Law e Lodge 1984:96)
[61] It follows that the sense of perceptual stability which we all routinely experience does not necessarily constitute evidence against the existence of a learned component in perception. (Law e Lodge 1984:96)
[62] Cues from authority will only be entertained and permanently accepted if they are workable: perceptual lumpiness must endorse what authority tells is to be seen. (Law e Lodge 1984:96)

LEARNING & PRACTICE (sensorymotor)
[63] Since classes group perceptual phenomena, learning must forge a link between perceptual lumpiness and class. It is intensely practical. (Law e Lodge 1984:96)
[64] Learning new classes depends upon extension from classes that are already practically usable. (Law e Lodge 1984:96)
[65] Thus certain classes may be ’empty’ without direct empirical referent, but still usable. (Law e Lodge 1984:96)
[66] Finally, scientific knowledge and learning is like all other knowledge. It too is intensely practical. (Law e Lodge 1984:96)
[67] The workability of knowledge is typically tested in the course of action. (Law e Lodge 1984:103)
[68] If knowledge is a guide to action, it must classify not only perceptual lumpiness in the outside world, but also perceptual lumpiness in the body and determine the relationship of the former with the latter. In addition it must include the ability to monitor and control internal perceptual lumpiness (action). (Law e Lodge 1984:103)
[69] There is thus a close relationship between perception and action. Action depends upon perception and may in turn make it possible. (Law e Lodge 1984:103)
[70] As in the case of simple perception, the provision of authoritative cues is important in the classification and association of perceptual lumps and actions. (Law e Lodge 1984:103)

OSTENTION (performance)
[71] It is often easier to demonstrate action rather than to describe it. (Law e Lodge 1984:103)
[72] In some cases it is not possible to describe a skill fully. Direct modelling may be necessary. Polanyi rightly stresses the role of direct modelling in the natural sciences with his notion of ‘tacit knowledge’, though it is certainly the case that some knowledge can be transmitted without such direct modelling. (Law e Lodge 1984:103)
[73] As is the case for perception, the only way of acquiring such a skill is through practice. (Law e Lodge 1984:103)

METAPHORICAL REDESCRIPTIONS
[74] In classifying – that is to say, in asserting a similarity – we treat new objects as if they were the same as old objects, even though this cannot be the case. (Law e Lodge 1984:118)
[75] Thus, all knowledge is metaphorical. It treats different objects as if they were identical. (Law e Lodge 1984:119)
[76] Metaphorical redescription involves the transfer from one domain to another, or from past to present experience, of a network of classes and their associations. It thus provides a model for thinking about events that have been so redescribed. (Law e Lodge 1984:119)
[77] Theory in science and social science is constitutively metaphorical, since it involves the extension of partially explicit networks of classes to new instances. (Law e Lodge 1984:119)
[78] The fact that knowledge in general and theory in particular is metaphorical is concealed where applications have become institutionalised. Under such circumstances people sometimes mistakenly believe that they have achieved literal description. (Law e Lodge 1984:119)
[79] Scientific reasoning, like all other, can be seen as metaphorical redescription. Metaphor is not unscientific. (Law e Lodge 1984:119)
[80] Thus science cannot be distinguished from non-scientific knowledge by virtue of an absence of metaphor. (Law e Lodge 1984:119)
[81] Metaphor (the application of a network) can be seen as a bridge which carries a two-way traffic. New experience alters the extension of the terms of the network and may influence their intension. This affects the application of those terms to previously described phenomena. (Law e Lodge 1984:119)

PREDICTION-CAUSALITY or EXPLANATION-PLAUSIBILITY
[82] Many, but not all, theories have predictive potential. (Law e Lodge 1984:119)
[83] In the social sciences in particular, many explanations are best seen as rationalisations. (Law e Lodge 1984:119)
[84] Explanations vary in their degree of generality. However, specific explanations depend on generally accepted assumptions about the relationship between classes. (Law e Lodge 1984:119)
[85] Styles of acceptable theory vary greatly between social structures. (Law e Lodge 1984:119)
[86] In natural science, theories are normally posed in an impersonal and causal mode. (Law e Lodge 1984:119)
[87] The plausibility of a theory depends heavily upon what passes as an acceptable mode of explanation in a given social structure. (Law e Lodge 1984:120)
[88] It follows that the evaluation of the workability of a theory is something that can only be done from within a given social context. The network theory itself is neutral between particular empirical theories. (Law e Lodge 1984:120)

KNOWLEDGE & INTEREST-BELIEF (natural accounting, social control & network alignment)
[89] Knowledge is constructed under the auspices of interests and may be seen as a resource designed to advance those interests. (Law e Lodge 1984:133)
[90] The growth of knowledge can only be analysed if the field of interests in terms of which it was developed is understood. (Law e Lodge 1984:133)
[91] Where actors or groups share interests, they have a tendency to align knowledge, and agree on the ‘facts of the matter’, as the same questions are being posed of the knowledge. Where interests are divergent there is a tendency to controversy. (Law e Lodge 1984:133)
[92] At the most general level, the relationship between social interests and knowledge is contingent. There is no necessary relationship between social position and belief. However, in given circumstances social interests become linked to particular belief systems as actors buy into the latter and use them to advance their interests. (Law e Lodge 1984:133)
[93] It is useful to hypothesise that knowledge is developed under the auspices of two interests: a practical interest in natural accounting, prediction or control; and a retrospective interest in social control or legitimation. (Law e Lodge 1984:133)
[94] Knowledge directed primarily by an interest in social control and legitimation normally achieves its aim in part by successful prediction, accounting, or control. Knowledge directed by a primary concern with prospective accounting has additional social control implications. (Law e Lodge 1984:133)
[95] The network/interest theory of knowledge suggests that science is not pure in the sense that it involves a disinterested search for truth. The search for knowledge is inevitably directed by interests. (Law e Lodge 1984:141)
[96] Often the interests are those of important audiences. Thus experts seek clienteles for their expertise and knowledge may be constructed by experts with the interests of those clients in mind. (Law e Lodge 1984:142)
[97] The operation of external social influences is not necessarily inimical to the production of workable empirical knowledge. (Law e Lodge 1984:142)
[98] Social interests may lead scientists to a detailed and entirely workable study of matters that had not previously attracted attention. (Law e Lodge 1984:142)
[99] Since belief is a function of prior learning and interests, it follows that arguments and ideas are powerless in and of themselves. They are only persuasive if they interact with prior belief and interest in a way that appeals to the hearer. (Law e Lodge 1984:153)

MATERIALISM
[100] Idealism is thus an unacceptable explanatory doctrine in the social sciences. Materialism, which suggests that people operate with and manipulate ideas (rather than vice versa) is the preferred mode. (Law e Lodge 1984:153)
[101] The rise and fall of knowledge is thus attributable to the fate of those who believe it and to their social context, rather than to any intrinsic power or weakness of the ideas themselves. The study of knowledge in this way is usually called the sociology of knowledge. (Law e Lodge 1984:153)

SOCIO-LOGIC (coherence conditions)
[102] What passes as a ‘logical’ argument is one that is in conformity with psychological and social coherence conditions. Since the latter are variable, what passes as a logical argument is also variable. (Law e Lodge 1984:153)
[103] Deduction cannot be logically justified. If people accept deductive arguments it is simply because they work that way. (Law e Lodge 1984:154)
[104] An argument can always be saved from a display of logical inconsistency by means of classificatory adjustment. Logic is thus a defeasible tool. (Law e Lodge 1984:154)

NORMAL SCIENCE SOCIALIZATION (network alignment)
[105] Practical network alignment is most likely where experience, social cues and interests are shared by all concerned. (Law e Lodge 1984:163)
[106] Science is an example of an institution which is designed to align relevant experience, social cues and interests, and hence to maximise the chance of network alignment. (Law e Lodge 1984:163)
[107] Scientific socialisation is a lengthy, rigorous, authooritative and practical process. It teaches students the proper ways of seeing, interpreting and understanding natural phenomena. (Law e Lodge 1984:163)
[108] Despite such shared socialisation, network alignment can only ever be achieved for practical purposes. This is because, in the face of anomalous empirical instances, the network itself offers no guide as to how it should be revised. This is, of course, a matter of choice. A decision has to be made. (Law e Lodge 1984:163)
[109] Continuing network alignment thus rests upon a continuing shared view about what constitutes the best decision, and hence on continuing shared interests. (Law e Lodge 1984:163)
[110] Practical alignment of networks occurs in certain areas of science – those areas that Kuhn calls ‘normal science’ – which fulfil the conditions described in proposition [105]. (Law e Lodge 1984:169)
[111] Normal science activity is in many ways similar to scientific socialisation. It involves a consensual attempt to bridge the gap between relatively theoret- ical and empirical parts of the appropriate network by manipulating both theory and data to achieve a workable match. (Law e Lodge 1984:169)

ANOMALIES (troublesome misfits)
[112] The fact of a past tradition of normal science does not guarantee that this will continue. Each new result may, in principle, be judged anomalous and lead to crisis or fragmentation. (Law e Lodge 1984:170)
[113] Any network will generate misfits, that is, instances that are not easily interpretable in terms of obvious network extension. This follows from the fact that a network necessarily involves loss of information. (Law e Lodge 1984:180)
[114] Whether such misfits are accorded significance and noticed as anomalies depends upon a decision by the actors concerned, which is in turn a function of their conceptual habits and social interests. (Law e Lodge 1984:180)
[115] Once a misfit has been accorded the status of an anomaly, there are further decisions that have to be made about how the anomaly will be handled. A range of options is available, though it should be understood that these overlap, or shade into one another. (Law e Lodge 1984:180)
[116] Where actors are strongly committed to their networks, there is a wide armoury of legitimations and excuses available which permits them to discount the apparent anomaly and preserve the network unaltered. This is an extremely common strategy. (Law e Lodge 1984:180)
[117] Sometimes, however, the anomaly is accorded significance. In this case a further set of possibilities is opened up. The anomaly may finally be reconciled with the network through the manipulation or revision of either or both the theoretical and empirical ends of the network. (Law e Lodge 1984:181)
[118] Another possible outcome is radical network change. In science this is sometimes called a ‘scientific revolution’, though there are always elements of continuity however radical the change (see proposition [53]). (Law e Lodge 1984:181)
[119] There is no easy or overall sense in which the new network may be said to be superior to that which preceded it. The best we can say is that it served the interests of the relevant actors more satisfactorily. (Law e Lodge 1984:181)
[120] The outcome of the interaction between potential anomaly and network, like standard network extension, is thus a function of social interests and conceptual habits. (Law e Lodge 1984:181)

PRIMITIVE BELIEFS & SOCIAL DIFFERENTIATION-COMPLEXITY
[121] If network theory is realistic, then both those beliefs that are thought to be true and those taken to be false should be explained by the same kinds of factors. We should, therefore, be ‘impartial’ with respect to truth and falsity and ‘symmetrical’ in explanatory approach. (Law e Lodge 1984:187)
[122] Primitive beliefs are as rationally based as our own. Primitives, like us, construct networks that adequately serve their practical interests. The psychological coherence conditions that underlie their networks are identical to ours. Accordingly we should explain their beliefs in the same way that we explain ours – impartially and symmetrically. (Law e Lodge 1984:199-200)
[123] Like scientists, primitives defend their networks by refusing to turn misfits into anomalies. (Law e Lodge 1984:200)
[124] Differences between scientific and primitive networks result from differences between the social contexts in which they are produced. Science is produced in social context that encourages certain kinds of change. (Law e Lodge 1984:200)
[125] Social differentiation tends to encourage network change. It does this for three reasons: (a) it may favour the active generation of esoteric knowledge; (b) it may foster a relatively non-protective attitude towards established belief; (c) it tends to provide opportunities for the generation of knowledge with few social control implications. (Law e Lodge 1984:200)

COMMON SENSE
[126] Like scientists, children and lay-adults operate in terms of networks. Though the content of these networks differs, overall they exhibit the same properties – the selection of similarity, difference, association and the rest. There is, in short, no ‘epistemological break’ between common sense and science. Impartial and symmetrical analysis of the two is appropriate. (Law e Lodge 1984:206)
[127] Common sense networks dictate their extension no more than do scientific networks. Events are susceptible to classification in more than one way, and where interests are divergent, negotiations about the ‘proper’ extension of a term may occur as actors seek applications that best accord with their own interests. (Law e Lodge 1984:206)

MARXISM & IDEOLOGY (Mannheim X Barnes)
[128] Marx offers a theory of ideology that rests in part upon a correspondence theory of knowledge, but also contains a strong pragmatic element. (Law e Lodge 1984:221)
[129] The correspondence theory relates to certain basic and purportedly real class-interests. The pragmatic element assumes that knowledge is a function of practice in a social structure. (Law e Lodge 1984:221)
[130] In so far as Marx explains the beliefs of groups in his theory of ideology, the explanation rests on the latter pragmatic theory and is accordingly in conformity with the tenets of impartiality and symmetry. (Law e Lodge 1984:221)
[131] Mannheim’s theory is similar to that of Marx, having elements of both a correspondence and a pragmatic theory of knowledge. The latter predominates. However, the distinction between ideology and ‘situationally congruous’ knowledge rests upon a modified correspondence theory of knowledge and thus breaches the tenets of impartiality and symmetry. Attributions of ‘ideology’ tend to become devoid of meaning in the pragmatic theory. (Law e Lodge 1984:221)
[132] Barnes’ theory – that of knowledge as a tool directed by two great interests, one an overt interest in prediction and control and the other a concealed interest in social legitimation – does not breach the tenets of impartiality and symmetry and is pragmatic in nature. (Law e Lodge 1984:221)
[133] However, all knowledge is directed in part by a concealed interest in social control, though this may coincide with an overt interest in prediction and control. Thus, to argue (as does Barnes) that certain knowledge – that directed by at least the former interest – is ideologically determined is unsatisfactory. (Law e Lodge 1984:222)
[134] To sustain an attribution of ideological determination it would be necessary to distinguish between local and global concerns with social control, and identify ideological determination with the latter. (Law e Lodge 1984:222)
[135] There seems to be little reason for making this distinction. It would, accordingly, be best to assume that terms such as ‘ideology’ and ‘ideological determination’ have no place in a symmetrical theory of knowledge. (Law e Lodge 1984:222)

RELATIVISM
[136] The network theory has relativistic implications because it notes that the success of argument and conceptions of what constitutes the truth can only be understood if the local circumstances of the production of knowledge are considered. There are, in other words, no general criteria for determining the truth or superiority of any network. (Law e Lodge 1984:228)
[137] Arguments against relativism from translation are not persuasive. Those that are based on claims about the universal nature of certain logical operations either fail to discriminate between languages or depend upon particular contextual judgements of the nature of these operations. Those that work from a version, however disguised, of the concept of a neutral observation language are inconsistent with the conception of observation found in network theory. (Law e Lodge 1984:228-9)
[138] The network theory is not relativistically self-defeating because it does not make final claims about its own correspondence with reality. (Law e Lodge 1984:229)
[139] The network theory does not suggest that knowledge is distorted by the operation of psychological or social influences. It notes, instead, that such influences direct the way in which empirically adequate networks are constructed. (Law e Lodge 1984:229)
[140] Network theory, in its relativism, does not deny or undermine intellectual standards. It simply notes that standards are locally generated. (Law e Lodge 1984:229)
[141] Nevertheless, ‘rationalism’ and relativism are incommensurable. Just as rationalism does not show that relativism is essentially flawed, so relativism cannot demonstrate that rationalism is wrong. (Law e Lodge 1984:229)

VERSTEHEN in NATURAL/SOCIAL SCIENCE (emic X ethic adequate causal stories)
[142] Members of a culture learn about social kinds in the same way as they acquire knowledge of natural kinds. Social institutions are acquired in this way, practically, from the interaction of ‘lumps’ in the environment and social cues. (Law e Lodge 1984:243)
[143] The act of verstehen involves the acquisition of a social network in this way. (Law e Lodge 1984:243)
[144] If it wishes to explain categories constituted in social networks, social science must to some extent use verstehen. (Law e Lodge 1984:243)
[145] However, the interests of social scientists differ from those of the members of the culture they are studying. Accordingly, what counts as an adequate account of a culture is a function in part of those interests, and that account or translation may properly differ from the native’s own version. (Law e Lodge 1984:243)
[146] In comparative analysis the social scientist treats as ‘the same’ parts of different cultures that, in principle, differ from one another. However, in doing so, he or she differs not at all from the natural scientist who also compares unlike with unlike. Accordingly, there is nothing wrong with comparative social science that is not wrong with comparative natural science. The problem in both cases is a practical one: is the comparison workable? (Law e Lodge 1984:243)
[147] The fact that links in networks of social explanation are weak is no reason for abandoning a language of causality. There are many weak links in networks of natural explanation. (Law e Lodge 1984:243)
[148] Since the interests of the social scientist differ from those of the people that he or she is studying, so too will their networks. Networks of the former, in seeking to account for actions of the latter may include some links or ‘causal stories’ which correspond to reasons available to the latter, and some which do not. There is no reason for the analyst to limit him or herself to those links available in native culture. Indeed, (see [149]) it is most unlikely that a satisfactory social-science explanation could do so. (Law e Lodge 1984:244)
[149] Adequate ‘causal stories’ are those that explain or account for an interesting, unusual. or otherwise noticeable event. They are explanatory because they re-describe the event in question by locating it in a different network. (Law e Lodge 1984:244)

SOCIAL SCIENCES as PUZZLE SOLVING
[150] The requirement sometimes imposed by social scientists that analysts’ explanations must be logically tied to those of the natives ignores the fact that what counts as a logical tie is negotiable and a function of interests. (Law e Lodge 1984:244)
[151] In the social sciences there are typically divergent views and understandings of phenomena of interest. (Law e Lodge 1984:254)
[152] This often arises because the analyses offered are directed at large non-specialist audiences, have major implications for social control and are subject to analysis from the standpoint of a variety of different interests. (Law e Lodge 1984:254)
[153] A unified social science would be facilitated in a society where all sponsoring interests were aligned – that is to say, a totalitarian society. (Law e Lodge 1984:254)
[154] Ecumenical attempts to unify social science by fitting together its components as in a jigsaw are likely to fail in the absence of an alignment of interests because they rest upon a correspondence theory of knowledge. (Law e Lodge 1984:254)
[155] A plausible way of working in social science is to select resources that have bearing on a limited range of puzzles and attempt to solve those puzzles while substantially ignoring alternative practices unless these can, in turn, be treated as resources. (Law e Lodge 1984:254)
[156] Successes will only be local, but they will at least be successes. (Law e Lodge 1984:254)
[157] This approach is not ‘irrational’. It rather notes that rationality or adequacy is something that can only be determined locally. (Law e Lodge 1984:254)

SOCIAL SCIENCE IS NOT PHILOSOPHY
[158] The network theory is not only descriptive. It also has prescriptive implications, suggesting, for instance, that agents are active, that ’empiricist’ and ‘ecumenical’ approaches to social science are likely to fail, and that science cannot be distinguished from social science in terms of its methods. (Law e Lodge 1984:262)
[159] The social sciences are not short of prescriptions, many of which come from philosophy. These are often presented in terms of an historical discussion of the development of philosophy. (Law e Lodge 1984:262)
[160] This type of prescription should not normally be taken seriously, because it comes from outside social science practice. There is no reason why social scientists should define their practice in terms congenial to the imperialist claims of another practice, that of philosophy. To do so is an unwarranted failure of nerve. (Law e Lodge 1984:262)
[161] Instead, social scientists should have confidence in their own locally defined practices, and simply treat philosophy as a resource like any other discipline. (Law e Lodge 1984:262)

LAW, John; LODGE, Peter. 1984. Science for social scientists. London: Macmillan Press.

Tags :