User:KYPark/2003/Samuels

A Companion to the History of Economic Thought

Warren J. Samuels, Jeff E. Biddle, John B. Davis

http://books.google.com/books?id=3H8gBQv5MysC

30 The Sociology of Economics and Scientific Knowledge, and the History of Economic Thought

A. W. Bob Coates

31 Exegesis, Hermeneutics, and Interpretation

Ross B. Emmett

32 Textuality and the History of Economics: Intention and Meaning

Vivienne Brown

32.1 Introduction

To say that studying the history of economics involves interpreting economics works and associated archival materials that were written in the past may seem uncontroversial, but such an apparently simple formulation of what is done in studying the history of economics is far from straightforward. Issues of interpretation have been the subject of intense debate over recent decades, and these debates have spanned (at least) philosophy, literary theory, history, and cultural studies. This essay will focus on just one aspect of these debates by engaging with arguments about the objective of reconstructing the author's intended meaning, and it will relate these arguments to the notion of "text" and "textuality." Given the high opportunity costs of engagement in such apparently arcane theoretical debate, economists may well wonder whether there is any particular potential payoff for their understanding of the history of their own discipline. A presupposition of this essay is that these theoretical debates do have something to contribute to the history of economics in helping to explain how different interpretations of the same works keep being produced, and so have implications for the ways in which the history of economics may be understood.

It is a commonsense presupposition that correctly interpreting a work gives us the meaning that its author intended; correctly reading Adam Smith's works, for example, gives us the meaning that Smith intended. To query this commonsense presupposition may seem to some to imply that we have already crossed the Rubicon that separates rational from irrational discourse, but two questions may be raised at this stage which should caution against such a hasty conclusion.

[...pp. 539-540 n/a...]

32.2 Intentionality

[...] of an illocutionary act requires that there is "uptake" of the act, in the sense that the audience or readership understands the intended illocutionary force of the utterance or work.

This approach has been developed in the work of Quentin Skinner in the history of political thought and has achieved a wide influence in intellectual history by this route (Skinner, 1988a-c; also critical articles in Tully, 1988). It is argued that to recover the historical meaning of a work it is necessary to recover the intended illocutionary force of the work in addition to understanding its sense and reference; or, alternatively, that to understand a work's sense and reference in historical context it is necessary also to recover its intended illocutionary force. To understand the historical meaning of a work it is therefore necessary to understand what the author was doing in writing such a work, and to achieve this it is necessary to recover the author's intentions in writing such a work, by placing it in its relevant historical context of argument and counter-argument as framed by the recognized linguistic conventions of the time. In an early statement of this position, it was argued that "the essential question which we therefore confront, in studying any given text, is what its author, in writing at the time he did for the audience he intended to address, could in practice have been intending to communicate" (Skinner, 1988a, p. 65). To answer this question requires the decoding of the author's intentions by placing that utterance in the context of the linguistic conventions of the time, which would have been taken for granted by the intended audience, and with respect to which the author's intentions in making that utterance would have been understood by those who were cognizant of those conventions. It was also argued in this paper that this historical approach requires that those intentions should be describable in terms that could have been accepted as correct by the author, even though the historian may be able to give a fuller or more convincing account of the author's intentions than the author could have given thus ruling out the use of later conceptual criteria that were not available to the author (1988a, p. 48).

To embed the history of ideas in the intentions of the author as an historical figure purports to give a decidedly historicist inflexion to the history of ideas, but this has been somewhat displaced in later statements of Skinner's position, where it is argued that an interpreter does not have to take the author's statements about the meaning of a work, since the author may be "self-deceiving" or "incompetent" in this respect (Skinner, 188b, p. 77). It is thus not the author but the interpreter who is the "final authority" concerning what the author was doing in a particular work )p. 77). This raises the question of the relationship between the intended illocutionary force and the actual illocutionary force of an utterance as interpreted by the historian. If establishing the actual illocutionary force of an utterance, or what is meant by an author's saying x, is, ultimately, not actual, not the intended, illocutionary force of an utterance that is taken to constitute the meaning of that utterance or what the author was actually doing in the . . .

[...pp. 542-4 n/a...]

. . . from their expressed ones, historians have to go beyond the expressed beliefs if they are to recover the actual ones" (Bevir, 1999, p. 267). But in this case the actual beliefs are those expressed by the work, not those expressed by the author (Bevir, 1999, p. 71). But how may the historian decide that the author's expressed beliefs are insincere, unconscious, or irrational? The injunction to the historian is to make sense of the material, since "the logic of their [the historians'] discipline remains such that they should invoke insincerity, the unconscious, or the irrational only if they cannot make sense of the material without doing so" (Bevir, 1999, p. 173), but the material that has to be made sense of here is not the expressed beliefs of the author but the works themselves. The reasons for deciding whether the author's expressed beliefs are sincere and so on are based on the historian's injunction to make sense of the material. The logic of the historian's discipline is thus that the work has to be made sense of by involving whatever authorial beliefs -- sincere or insincere, conscious or unconscious, rational or irrational -- are thought necessary to achieve this (Brown, 2002). Bevir's weak intentionalism thus requires attributing beliefs to authors which they may not have been aware of having, or which they might even have rejected as not their own, and in this respect his account is similar to Skinner's. In casting the author's pre-conscious and unconscious intentions as intentions none the less, the historian is required to provide a rationalized reconstruction of the author that relies on some modern understanding of psychoanalytic theory, and which may not correspond with the self-consciousness of the actual historical figure, even though it is the historical experience of that figure that ostensibly ensures that the intended meaning is an historical meaning. As with Skinner's and Hirsch's accounts, what drives the interpretation of the work is the need to attend to the actual beliefs (or actual illocutionary force or actual implications) as interpreted in the work under study, and any imputation of author's meaning is an inference based on that interpretive requirement. What is missing in such talk of the dependence of the work's meaning on the intentions of the author is any works (and other archival materials) that are being proposed by the historian. This is another instance of the difficulty faced by Skinner's and Hirsch's approach, that there is but one entity -- the works and other archives in question -- and so there is no evidence of the author's conscious and unconscious intentions independently of the interpretations being offered of the works, and which could be called upon to assess the extent to which any particular interpretation is a valid reconstruction of those intentions (cf., Hirsch, 167, p. 165). Bevir's argument that objectivity is achieved by comparing rival interpretations with respect to various criteria does not meet this point, since what are being compared -- according to his account -- are rival interpretations of the works, not rival interpretations of the author's expressed beliefs.

This section has considered examples of the three different categories of authorial intentions that are presented in Davidson's passage, and it has argued that none of these provides a coherent account of the argument that the meaning of works is given by the (conscious and unconscious) intentions of their authors. This is not to say that accounts of authorial intentions would not be interesting or worth having if they could be found. The difficulty that is common to such approaches is (at least) twofold. First, there is the issue of the evidence of what those intentions were. If such evidence does not exist independently of the interpretations of the works and other archives that are at issue, then there is no independent evidence in principle relating to the author's intentions -- as opposed to the evidence relating to the interpretation of the works. This implies that such claims about providing knowledge of the author's intentions are epistemologically empty. Secondly, imputing particular intentions to the author on the basis of an interpretation of the works may involve an interpretive reconstruction of the author's intentions, which the author as historical agent neither would nor could have assented to if they were put to him or her. This implies that such reconstructions of the author's alleged intentions cannot make claim to be invoking the historical author's intentions as they existed in time for him or her as a self-conscious agent, and so the strictly historical significance of such interpretations is open to question. What remains of such claims to have discovered or recovered the author's intentions thus amounts either to a rhetorical (in the narrow sense of merely persuasive) gesture, or a complimentary gesture bestowed upon valued interpretations which are simply supposed to have successfully reconstructed the author's intentions (cf., Ankersmit, 2000, pp. 325-6).

32.3 Meaning

The intentionalist approaches reviewed in section 32.2 have in common the argument that the meaning of a work is given by the psychological or mental state (or content of the mental state) of the author, whether this is constructed in terms of the author's intentions, determining will, beliefs, or viewpoint. The meaning of the work is thus determined by some psychological or mental entity that exists in a relation of exteriority to the works but yet is held to determine the proper interpretation of those works. This meaning, determined externally, is then somehow intrinsic to the work and so can be "grasped," "found," "uncovered," or "recovered" in the act of interpretation. It is this that explains the tension in intentionalist arguments between the exteriority of the source of meaning with respect to the author's intention and the interiority of that same meaning with respect to the work itself. Caught between these two alternative sites for the definitive center of a work's meaning, these intentionalist arguments inevitably have to recognize that it is indeed the latter that is definitive but, as argued above, this renders precarious the notion of author's intention.

There is one version of the intentionalist argument, however, that does not encounter this problem. This version identifies the author's intended meaning wth the interpretation of the work such that what emerges from a process of interpretation just is the author's intended meaning, since that just is what interpretation delivers (Knapp and Michaels, 1985' for a discussion, see Mitschell, 1985). Instead of positing the author's intended meaning as something which has or did have an independent existence vis-a-vis the work, the author's intended meaning is by definition held to be that which an interpretation renders. This version of the intentionalist argument overcomes the problems outlined in the previous section, since there is no longer any claim that the author's intention had a historical existence independently of interpretations of the work or that it is something that the author as a historical figure could have recognized or identified as his or her own. Such an argument is therefore not subject to the criticisms advanced in section 32.2, but this is achieved by making the argument tautologous. If any interpretation is by definition a rendering of the author's intended meaning, then the significance of its being so is surely lost. If its being the author's intended meaning no longer operates as a constraint on what may be claimed in an interpretation (over and above what the evidence of the work can sustain), then surely there is no longer any point in claiming it to be so.

According to Hirsch's argument, once the assumption of an author's intended meaning is removed, then no interpretation can correspond to the meaning of the text, because the text itself has no determinate meaning and can say different things to different readers (Hirsch, 1967, pp. 5, 11). Similarly, according to Bevir's argument, meanings can not be ascribed to texts in themselves; unless the interpreter specifies for whom the text had that particular meaning, then the interpreter is merely saying how he or she chooses to read the text, and in this case "we should not make the mistake of assuming he is offering us an interpretation of the text itself; we might enjoy his reading, but we should not bother to ask ourselves whether it is true or not for there is no object of which it seeks to give an adequate account" (Bevir, 2000, p. 391). If the notion of author's intended meaning according to these intentionalist arguments is not accepted, however, this implies that all there can be are meanings that are the product of different readings. According to intentionalist arguments such as Hirsch's and Bevir's, such readings suffer the lack of not offering an account of the author's meaning, but if the author's meaning (as something other than an interpretation of the work) is not accessible, then the notion of a reading is a more coherent notion than author's meaning. What is construed as a weakness from within an intentionalist argument is thus regarded as a strength once that intentionalist argument is recognized as deeply problematic. Furthermore, it is not the case that there is no "object" of which a reading seeks to give n adequate account, since that object is provided by the text/work. Indeed, even within intentionalist accounts this has to be the case, since there is no object apart from the works of which an account can be given. As argued above, even according to the intentionalist accounts of Skinner, the historian seeks to give an account of the actual illocutionary force; for Hirsch, the interpreter seeks to construe the implications of the text; and for Bevir, the historian of ideas seeks to make sense of the materials. Inevitably, all that intentionalist interpretations can ever hope to achieve is to provide readings of the texts/works, and this is made evident in their own writings.

The differences between these various positions may be illustrated by reference to the issue of the "symmetry" or "asymmetry" of information between the analyst and the agent being studied (e.g., Sent, 1998). In the case of interpreting works in intellectual history, the analyst is the intellectual historian and the agent is the author of the work. Is there symmetry of information between the historian and the author? The author presumably has some access to the contents of his or her own mind that is denied to the historian. In addition, the author's information about the biographical, historical, and intellectual circumstances of the writing of the work is different from the historian's information about these topics. Both may have equal information about the contents of the works, however, on the assumption that there are no significant problems with missing or corrupted editions of the work (an assumption that is not always appropriate). Assuming that the historian and author have the same information about the contents of the work, there seems to be a major asymmetry in terms of knowledge of the author's mind and knowledge of the circumstances of writing the work. How do the different approaches to the relation between intention and meaning discussed above deal with this? Hirsch's argument provides discussion of the ways in which the interpreter may gain knowledge of the author's intended meaning but concedes that, in the end, interpretation amounts to refined guesswork. This suggest that the asymmetry is such that the historian's information is inferior to the author's. Skinner's and Bevir's arguments, however, reverse this asymmetry; each one privileges the interpreter's information, as the interpreter may have a better understanding of the author's intentions than the author has him- or herself, whether this superior information is based on the superiority of modern historical research or on modern psychoanalytic theory. The argument put by Knapp and Michaels, however, collapses the information relation between interpreter and author into one of identity; as an interpretation of the work just is the author's meaning, there can be no difference in principle between the two. The argument of this essay, by contrast, is that the asymmetry in information between interpreter and author is insuperable, as the interpreter cannot gain access to the mental state of authors who are now deceased. The proper object of interpretation is the work itself and not an inaccessible mental state, and so the issues raised by a/symmetry of information do not apply.

32.4 Textuality

There are various senses in which a piece of writing may be said to be a "text." The word "text" itself is the past participle stem of the Latin verb texere, to weave, intertwine, plait, or (of writing) compose. The English word "textile" and "texture" also derive from the same Latin word. this etymology of the word "text" is apparent in expressions that refer to the "weaving" of a story, the "thread" of an argument, or the "texture" of a piece of writing. A "text" may thus be taken to be a weaving or a network of analytic, conceptual, logical, and theoretical relations that is woven with the threads of language. This implies that language is not a transparent medium through which arguments are expressed, the invisible and self-effacing carrier of a message from the mind of the author to the mind of the reader, but is interwoven with or provides the very filaments of the substantive arguments themselves. Construing a piece of writing as a "text" in this sense thus foregrounds the issue of the relation between the weaving of the argument and the linguistic threads and filaments by means of which the text is woven. Issues of language -- style, figuration, rhetoric, the polysemy of words -- are thus not extraneous to the meaning ascribed to the text in the process of reading but are a constitutive part of it. Reading a text thus requires a practice of "close reading" that examines the woven network of textual relations and how the substantive, logical, and linguistic aspects of the text may be intertwined. This is uncontroversial for overtly literary writings, but its implications are not always taken into account for nonliterary writings (for an attempt at such a reading of Adam Smith's works, see Brown, 1994, 1997b).

In examining the texture of a piece of writing, it is thus an issue for investigation as to what kinds of linguistic threads there are and how they might be interwoven. The various threads may well be drawn from different "language," such that different historical moments may be interlaced with different analytic, conceptual, and axiological systems. An issue for investigation is the extent to which a text may be an ordered text with a determining structure, logic, or language, or whether it is a tissue (also deriving from texere) of heterogeneous languages and hence internally differentiated or textured. Also, there is the question of whether and to what extent the text is laid open, smooth, and without creases or marks, or whether it is folded in upon itself so that the patterning or texture is subject to breaks or introversions. Furthermore, the text may not have finished or clearly defined edges, since the threads may be left hanging loose or interwoven with other texts.

Texts function within contexts of interpretation. to suggest that a text needs to be read in terms of a context is, however, not to prescribe a single method of reading but, rather, to open up some of the issues that might be involved in any particular method. To suggest that the context should be a historical one, in order to generate a historical reading, is also not to prescribe a single kind of reading, since there may be no single authentic historical moment that coincides with the originary moment of the creation of the text and which determines a single appropriate historical context for reading it. If a given text incorporates different historical threads, it is thus still an issue as to how these threads are to be followed in pursuing a reading. Such threads may also relate in different ways to what is known (if at all) of the biography of the stated intentions of the author, or of the economic, social, political, philosophic, theological, artistic, cultural, or intellectual motifs of the time. As more is learnt about the possible historical contexts that may assist a reading of a text so there may be a plurality of possible contexts -- different interweavings or different forms of intertextuality -- and an increase in the possible readings of a work which may be justified by the textual evidence. This plurality may well sit uneasily with a single discipline's conception of the canonic status of an exemplary author's works, as such disciplinary readings may tend to promote a unilinear or monologic notion of the work even though such works may still be liable to competing interpretations within the discipline. A discipline's "conversation" with its own canonic authors may thus turn out to be more dialogical, more multi-voiced, more open-ended, and less discipline-based, than canonic presuppositions might suggest (Brown, 1993). (The argument of these paragraphs derives from a large literature, which includes writings by Bakhtin, Barthes, De Man, Derrida, Fish, Foucault, and LaCapra.)

The notion of a "text" as a woven network of arguments and language is, however, itself only a metaphor. Its usefulness derives from the way in which it focuses on the need for a close reading of argument and language (which need not exclude attention to conceptual clarity, logical rigor, and analytic coherence). Exploring the resources of the text thus helps to prevent the abuse of claiming an unknowable authorial intention to buttress particular interpretations. But there are also limits to the applicability of this metaphor. The study of a woven cloth is conducted in a medium that is different from the medium of the cloth; in the case of written texts, the study is conducted in the same medium as the text itself. There is thus no meta-language for analyzing the language and argument of a text' and each interpreter is in turn an author, each interpretation in turn another work.

Any work, any piece of connected writing, however, is a self-referential entity. Words and expressions refer backward and forward; an argument makes a point, leads into another or draws to a conclusion; and narrative structures provide commentary or a sense of direction on what is being argued or concluded. The language of a piece of writing thus seems to instantiate intentional agency in that it is self-referentially engaged in doing something all the while; and this helps to explain how a piece of writing seems to have its own "voice" or "persona," even when that voice or persona is being disrupted or displaced by the workings of the language of the piece. This may be illustrated by the ways in which purposive argument is ascribed to a piece of writing. Instead of saying that author A meant such-and-such, or that author A argued such-and-such, it could be said (as in this essay) that the work or text says such-and-such, or that it argues such-and-such, thereby ascribing purposive argument to the work or text rather than to the author. Expressing the arguments of the text in the passive voice (such as "it is argued that" or "the argument is offered that") does not seem to eliminate this implied agency either, since the actions of "arguing" and "offering an argument" --even when expressed in the passive voice still seems to register the "voice" of an argumentative agent. This suggests that a form of agency is still being registered as present to the text even when the interpretation being offered of the text is not one cast in terms of the author's intentions. Intentionality of language for the intention of a particular author of an individual work. In other words, that intentionality is registered in language as a signifying system does not imply that the meaning of any particular piece of writing is determined by the intentions of the author of that piece. Sections 32.2 and 32.3 have outlined some of the difficulties involved in sustaining intentionalist arguments, and this section has tried to present an alternative approach to interpretation or reading. The rejection of authorial intended meaning as determining the meaning of a text is thus not a criticism of all notions of agency or intentionality, but represents an alternative way of trying to understand the ways in which complex notions of agency are inscribed within language, irrespective of, or even contrary to, the intentions of individual writing subjects. In this sense, the structures of language do not simply mirror the ways of the self, but form part of the context in which complex notions of the self are negotiated, refracted, contested, and sublimated.

32.5 Conclusion

In the Introduction to this essay, two issues of interpretation were raised: To what extent is it to be expected that the history of economics should be characterized by a high or increasing degree of consensus, and what are the procedures and criteria for assessing the relative validity of different interpretations? These are large issues that can hardly be addressed satisfactorily in a conclusion. To the extent that the assumption that interpretation seeks the recovery of the author's intended meaning suggests that progress in the history of economics ought to be associated with some sort of convergence of interpretations, the relinquishment of that assumption might provide some reason why such convergence is not necessarily to be expected, and so might provide an explanation of the increasing range of interpretations in the history of economics. But such a relinquishment does not imply a loose conception that "anything goes." On the contrary, the emphasis on close readings suggests that textual analysis should become more -- not less -- rigorous, and that increasingly high standards of textual evidence should be expected in defense of any interpretation. And on the question of the appropriate evidence for an interpretation, there is no evidence of the author's intention that is independent of the interpretation being offered of the works. This suggests that intentionalist interpretation cannot in practice be sustained.

Note

I would like to thank the Editors for their comments on an earlier draft of this essay. I am especially indebted to John Davis for raising the issue of symmetry of information.

Bibliography

F.R. Ankersmit. 2000. Comments on Bevir's The Logic of the History of Ideas. Rethinking History, 4, 321-31.

J.L. Austin. 1975 [1962]. How to Do things with Words.

M. Bevir. 1999. The Logic of the History of Ideas. Cambridge University Press.

Vivienne Brown. 1994. Adam Smith's Discourse: Canonicity, Commerce and Conscience. Routledge, London.