人大经济论坛

 找回密码
 立即注册

QQ登录

只需一步,快速开始

搜索
热搜: 活动 交友 discuz
12
返回列表 发新帖
楼主: 布拉格之春

Critical reflections on a realist interpretation

[复制链接]
 楼主| 发表于 2017-9-12 22:36:33 | 显示全部楼层
3.2. Confusion over ‘unrealisticness’

As Mäki points out, the strong instrumentalist version of F53’s torso (viz., that unrealistic assumptions can be desirable) is a popular one among economists. Mäki, however, asserts that, as a methodological doctrine of science, it is obviously absurd. By implicitly assuming that Friedman could not have been so confused as to have believed such a wrongheaded doctrine, it should be judged to be a highly implausible interpretation of F53. This provides an opening for Mäki’s third version of the torso: that assumptions, although inevitably unrealistic, should nonetheless be realistic enough – that is, sufficiently good approximations of reality – for one’s purposes.

Before examining Mäki’s third version of F53’s torso, we should first examine the strong instrumentalist version a little more closely because, if this interpretation is not ‘obviously’ a misreading by F53’s own lights, then the basis for appealing to a third, new version of the torso is somewhat undercut.

Let us begin by recalling that F53 distinguishes between the ‘validity’ and the ‘importance’ (synonymously, ‘significance’) of a theory. For F53, there is one clear criterion for judging a hypothesis’s or theory’s validity; viz., its ‘predictive power for the class of phenomena which it is intended to “explain”’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
; p. 8); ‘the only relevant test of the validity of a hypothesis is comparison of its predictions with experience’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
; p. 9). This criterion enables easy judgements of relative worth when comparing rival theories ranging over the same domain, where one is more predictively accurate than others. At this point, nothing is implied about the unrealisticness or otherwise of a theory’s various conceptual assumptions. F53 recognises, however, that generally speaking, validity is insufficient for choosing between rivals. Thus, additional criteria for theory choice are necessary. F53 nominates ‘simplicity’ and ‘fruitfulness’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, pp. 9–10).

Of these, the former is particularly relevant to the strong instrumentalist version of F53’s torso. F53’s definition of simplicity is this: ‘a theory is “simpler” the less the initial knowledge needed to make a prediction within a given field of phenomena’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 10, emphasis added). It is assumed here that the ‘initial knowledge’ being referred to is empirical knowledge of the agents and institutional circumstances that are referred to in a theory.1414. Since ‘initial knowledge’ is taken to refer here to empirical knowledge, it does not refer to what might be called non-empirical knowledge, such as knowledge of the rules of logic or Euclidean geometry or calculus. As such, a theory could still be deemed ‘simple’ in terms of requiring very little prior empirical information to make predictions, whilst simultaneously requiring substantial prior knowledge of, say, the mathematical operations necessary to construct and solve equations that contain that small amount of empirical information.
View all notes
Here then we come to the matter of unrealisticness. Greater simplicity of a theory (in the above sense) entails greater abstraction from reality – one strips away more and more representational detail – so that less and less prior empirical knowledge is required for a theory to generate predictions of interest. Greater unrealisticness of the theory, for F53, thus amounts to greater simplicity of the theory by means of abstraction from ‘the mass of complex and detailed circumstances surrounding the phenomena to be explained’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 14).1515. Note that although Mäki has long argued that abstraction and idealisation can be understood in ways that cohere with ‘realistic’ assertions (e.g. Mäki, 1994 Mäki, U. (1994). Reorienting the assumptions issue. In R. Backhouse (Ed.), New directions in economic methodology (pp. 236–256). London: Routledge.
[Crossref], [Google Scholar]
, 2009c Mäki, U. (2009c). MISSing the world. Models as isolations and credible surrogate systems. Erkenntnis, 70, 29–43.10.1007/s10670-008-9135-9
[Crossref], [Web of Science ®], [Google Scholar]
), the issue here is how F53 alone conceptualises ‘unrealistic’.
View all notes
For example, a theory aimed at predicting changes in world oil prices which, abstracts from empirical knowledge of, say, the actual aims and decision-making processes of oil suppliers, the actual technical capacities of and constraints on oil extraction in different locations, and Middle East geopolitics, would be a ‘simpler’ theory than one that attempted to integrate descriptively accurate representations of these factors. The theory would be more ‘unrealistic’ than the alternative just in the sense that it abstracts from some of ‘the reality’ of actual oil market participants and the technical and institutional contexts in which they operate.

Theoretical simplicity (and thus unrealisticness in F53’s sense), combined with actuated fruitfulness, then, gives us the criterion by which to judge the relative importance or significance of a theory:

Truly important and significant hypotheses will be found to have ‘assumptions’ that are wildly inaccurate descriptive representations of reality, and, in general, the more significant the theory, the more unrealistic the assumptions (in this sense). The reason is simple. A hypothesis is important if it ‘explains’ much by little, that is, if it abstracts the common and crucial elements from the mass of complex and detailed circumstances surrounding the phenomena to be explained and permits valid predictions on the basis of them alone. To be important, therefore, a hypothesis must be descriptively false in its assumptions; it takes account of, and accounts for, none of the many other attendant circumstances, since its very success shows them to be irrelevant for the phenomena to be explained. (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 14)

So we say that a theory is important if it ‘“explains” much by little’ where ‘much’ refers to the precision and scope of confirmed predictions and ‘little’ refers to the empirical knowledge included in the theory’s assumptions. We would then measure a theory’s importance by the ratio of precision and scope of confirmed predictions to the simplicity of assumptions. Ceteris paribus, the greater the simplicity of a valid theory, the more important it becomes. Thus, at the limit, it is entirely possible, by F53’s lights, for a theory (aside from its predictions) to be entirely empirically false and highly virtuous with respect to its simplicity. As an aside, it should be noted that simplicity here does not refer to the number of assumptions; a theory may have as many assumptions as one likes. Again, what makes the theory ‘simple’ is that these assumptions are not as dependent on prior empirical knowledge. They may, for example, be pure fictions, mathematical devices, fanciful metaphors or whatever, of which one would have to acquire knowledge – but this would not be empirical knowledge of the agents and circumstances to which a theory might be applied. So, when comparing rivals which are otherwise identical with respect to validity and fruitfulness, the most fictionalised theory – the one least dependent on prior empirical knowledge for its assumptions, and so is ‘simplest’ and most ‘unrealistic’ – should be judged the most ‘important’ of the batch. Thus, it would seem the strong instrumentalist version of F53’s torso can be faithfully derived from the text.

There is a potential fly in the ointment for the strong instrumentalist version of F53’s torso, however, because a maximally simple (say, maximally fictionalised) valid theory should, according to F53, contain ‘common and crucial elements’ required to ‘predict’ and ‘explain’ the phenomenon of interest. If the strong version is correct, that may sound strange or just muddle-headed, but this is so only if one assumes the phrase ‘common and crucial elements’ is referring to reality per se rather than just to whatever theoretical devices enable empirically parsimonious predictions of interest. It is not sufficient to assert that the strong version (as elaborated here) must be an incorrect interpretation of F53 just because this version has no specifically realist epistemic virtues. If the strong version is taken seriously, then it has built-in methodological virtues; the chief one being epistemic efficiency: the generation of correct predictions (epistemic output) with less initial knowledge (epistemic input) than demanded by realism.

Mäki seeks to counter such a move with a seeming reductio ad absurdum:

The strong version suggests that there might be even better theories that assume that air pressure is infinitely large and that businessmen aim at maximising their losses – these assumptions would be more unrealistic than the ordinary ones. But obviously, such unrealistic assumptions would not be epistemically virtuous, thus the strong version is questionable. (p. 95)

But this reductio mischaracterises the strong version (outlined above). The strong version does not just say any kind of crazy unrealistic assumption would make for a more virtuous theory. It says that if theory predicts well and is simpler (and thus more unrealistic), then that greater simplicity is a virtue. In addition to this, Mäki’s reductio implicitly rhetorically trades on the fact that the reader is likely to believe that the assumptions ‘air pressure is infinitely large and that businessmen aim at maximising their losses’ will be predictively unsuccessful (rather than successful, which is a condition of the strong version). If we were to take a less familiar example, however, the seeming reductio is rendered problematic. Let’s say we have a theory which says that male Australian redback spiders possess an instinctual urge for self-preservation. Paraphrasing Mäki: the strong version suggests that there might be a ‘better’ theory, which assumes, more unrealistically, that say, male Australian redbacks are gripped by a Freudian death wish. Is it now ‘obvious’ that ‘such unrealistic assumptions would not be epistemically virtuous’? According to the strong version, if it turned out that the death wish theory was more predictively successful than its rival for the domain of interest, the theory would indeed have greater epistemic virtue.16
 楼主| 发表于 2017-9-12 22:36:49 | 显示全部楼层
3.3. Clarifying ‘unrealisticness’

Let us leave aside the issue of whether the strong instrumentalist version of F53’s torso makes sense within the context of F53, and turn to Mäki’s alternate version, viz., assumptions should be ‘sufficiently good approximations for the purpose in hand’. Mäki assumes this third version is clearly realist in a manner that makes a difference vis-à-vis instrumentalism. But, this too is questionable because it is arguable that F53’s references to ‘sufficiently good approximation’ are, methodologically speaking, fairly inconsequential.

It is important to note that F53 takes the attribution of ‘sufficiently good approximation’1717. This includes variations such as ‘realistic “enough”’ and ‘sufficiently close’.
View all notes
to be a judgement that is made in the post-testing phase, and that ‘testing’ (in its preferred or ideal form) is predictive testing against rivals (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
; pp. 15, 16–17, 41). In the pre-testing phase, the question of the ‘actual degree of realisticness’ can de jure be ignored because according to F53 there is no way of knowing whether a proposition should count as a ‘sufficiently good approximation’ before predictive testing has occurred. Indeed, this is the nub of F53’s objection to those who attack orthodox theory for its unrealistic assumptions (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, pp. 30–31). So assuming the phenomenon of interest P is given, and the ‘virtues’ of simplicity and fruitfulness are held constant and equal for both theory T1 and its rivals T2 … n, we have:

(a) If T1 is more predictively successful than T2 … n for P then T1 is judged to be a sufficiently good approximation of reality for P.

Indeed, we can go further. As far as can be told, F53 never suggests or gives examples of the possibility of a theory being predictively unsuccessful (invalid) and simultaneously judged a sufficiently good approximation of reality. This being so, it may be reasonable to infer that the antecedent in (a) is in fact a necessary and sufficient condition of the consequent. As such, we may rephrase (a) as a biconditional.1818. Letting S = predictively successful and G = judged sufficiently good approximation, if it were the case that F53 was stipulating only ‘S→G’, then ‘~S & G’ should be possible. If, however, the stipulation is intended to be ‘G↔S’, then ‘~S & G’ is not possible (as seems to be implied by the text).
View all notes
Again, assuming given P and constant and equal virtues of simplicity and fruitfulness for T1 and T2 … n, we would then have:

(b) T1 is judged to be a sufficiently good approximation of reality for P iff T1 is more predictively successful than T2 … n for P.

So, whatever the truth values assigned to T1’s assumptions, however those assumptions are arrived at, and whatever the rationale is behind making those assumptions, if T1 implies predictions for P that are better empirically confirmed than its rivals, then T1 would be judged to be ‘realistic “enough”’ and its assumptions would be judged ‘sufficiently good approximations for the purpose in hand’. The question of the ‘actual degree of realisticness’ of assumptions effectively amounts to an automatic inference, and methodologically a mere addendum, in the post-testing phase.

So when Mäki says that pragmatic considerations enter into judgements of realisticness (p. 95), it should be recognised that these judgements (such that they are) are, by fiat, parasitic upon relative predictive success – that is, parasitic on the predictions being deemed ‘sufficiently accurate’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 15). And when Mäki says that ontological considerations of causal power enter into the judgement of realisticness (p. 95) (charitably assuming that they do in fact enter into such judgements for F53), this too is parasitic on relative predictive success – that is, there is an inference from relative predictive success to a subsequent ontological claim. Again, this is a post-testing judgement. In the pre-testing phase there is no need to give consideration to ‘realisticness’ of ‘core’ assumptions or the theory overall; and in the post-testing phase (in the unambiguous cases1919. That is, in cases where there is no conflict between the ‘virtues’, such as a simpler theory versus a more fruitful theory where both are equally predictively successful.
View all notes
) such considerations are really just cases of ‘rule-following’. Certainly, the choice to follow the rule may be a matter of philosophical judgement, but that is not a concern F53 expresses. For F53, following ‘rule’ (b) is presented quite dogmatically: it is just how F53 characterises positive science. One accepts the rule or one doesn’t, and that is that.

At a pinch, one could call the inference to an ontological conclusion ‘realist’ in the sense that the inference is at least about a theory-independent ‘reality’. But, this kind of realism is exceedingly permissive. A robust kind of scientific realism that is more deserving of the title is one that is concerned first and foremost (but not exclusively) with expressing and establishing a theoretical framework that first and foremost posits actually existing causal mechanisms (Bhaskar, 1975 Bhaskar, R. (1975). A realist theory of science. Leeds: Leeds Books.
[Google Scholar]
; Lawson, 1997 Lawson, T. (1997). Economics and reality. London: Routledge.10.4324/SE0085
[Crossref], [Google Scholar]
), capacities of entities (Cartwright, 1999 Cartwright, N. (1999). The dappled world: A study of the boundaries of science. Cambridge: Cambridge University Press. 10.1017/CBO9781139167093
[Crossref], [Google Scholar]
), dispositional properties (Ellis, 2002 Ellis, B. (2002). The philosophy of nature: A guide to the new essentialism. Chesham: Acumen.
[Google Scholar]
), natural kinds (Sankey, 2008 Sankey, H. (2008). Scientific realism and the rationality of science. Aldershot: Ashgate.
[Google Scholar]
) and the like that would serve to explain phenomena of interest. F53’s permissive (so-called) realism, by contrast, is one in which what counts as ‘realistic “enough”’ depends on whether a theory is a ‘good enough’ tool for predicting facts of interest to the investigator. This is effectively instrumentalism with an ex post realist appendix. If that is all realism amounts to, then with sufficient hubris and imagination almost any reformation of the methodological criterion in (b) could be given, and it could still count as ‘realist’. For example, take the following:

(c) T1 is judged to be a sufficiently good approximation of reality for P iff T1 is more mathematically complete [… or more inter-theoretically coherent or more useful to immediate policy formation …] than T2 … n for P.

Here the so-called realist element (‘sufficiently good approximation of reality’) is actually quite superfluous to making a judgement about the ‘worth’ of T1. It is an outcome of a judgement, not an input into one. The real ‘work’ is being done by the test criterion, whatever it happens to be (predictive power, mathematical completeness, inter-theoretical coherence, etc.). One could drop references to ‘sufficiently good approximation of reality’ and nothing would change vis-à-vis deliberations over theory choice.

It is possible that Mäki has been misled by a passage in F53 that is certainly suggestive of a kind of causal realism:

Why is it more ‘unrealistic’ in analyzing business behaviour to neglect the magnitude of businessmen’s costs than the color of their eyes? The obvious answer is because the first makes more difference to business behaviour than the second. (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 33)

Mäki interprets this passage as showing that causal realism ‘matters a great deal. It is advisable to consider costs – but not the eye color – because they make a difference for the phenomena being explained’ (p. 96). But here Mäki is treating F53’s shortened account of what is going on as though it were a complete statement. Why do we say that costs ‘make a difference’? It is because of ‘the effect on the discrepancy between actual and predicted behaviour of taking the one factor or is it the other into account’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 33). Now, which is it that matters here for F53? Is it the different predictive outcomes due to abstracting from eye colour versus costs, or is it the apparent subsequent, post-test, ontological inference that different costs cause different behaviour whereas different eye colour does not? The former is what really matters for F53. By leaving out eye colour and including an assumption about costs, one obtains better predictions and so one can now (post-test) say that the exclusive assumption about costs is more ‘realistic’ and a ‘sufficiently good approximation’ of (causal) reality for our purposes.2020. Although F53 is setting up an obviously uneven contest here, the example is not one of literally ‘no contest’. One could formulate, say, a racist eye colour theory of business behaviour that could generate predictions intended to rival a cost theory. F53, in addressing economists, is simply taking it as obvious that the cost theory (whatever the functional form for costs used) would be the more predictively successful one. From that, the judgement of ‘sufficiently good’ follows.
View all notes
Note that this example does not require that one represent cost functions in an empirically accurate way. Nor does it require that one posit any actual causal mechanisms relating to actual decision-making processes or institutional arrangements. One could theorise, say, positive and increasing marginal cost functions and a legally unregulated market, in direct contradiction of what is empirically known of reality in many cases, and this could still be a ‘sufficiently good approximation’ of reality simply because it predicts better than an eye colour theory. Answers to the question of why this is so, is an optional extra that, from F53’s perspective, does not bear on judgements about the theory’s validity and acceptability.
 楼主| 发表于 2017-9-12 22:37:03 | 显示全部楼层
3.4. Further confusion and clarification

Mäki’s point about the distinction between two different conceptions of unrealisticness – incomplete true descriptions of all reality and false descriptions of real particulars – is well taken, as is his point that abstracted or idealised ‘unrealistic’ assumptions can be compatible with realism. But being compatible is not the same as being constitutive; abstraction and idealisation do not in themselves lead us inescapably to a realist stance. More is required than that.

Most expressions of scientific realism are aspirational and prescriptive: for example, theories should, as far as is possible, refer to real entities and real causal relations that explain real phenomena; and where reference claims can be warranted, such theories should be believed to be true (e.g. Devitt, 1991 Devitt, M. (1991). Realism and truth (2nd ed.). Oxford: Blackwell.
[Google Scholar]
). Mäki needs to establish that F53 claims at least some ‘core’ assumptions should be true, and that overall, a validated parsimonious theory should be believed to be true. Alas, no such clear statement is forthcoming from F53. Indeed, F53 conspicuously avoids talking about ‘belief’ in the ‘truth’ of assumptions or theories. F53 instead uses much weaker terminology: one may accept a theory as valid and have confidence in its future validity (e.g. Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, pp. 8, 9, 18, 21, 22, 27). When it is recalled that ‘validity’ in F53 just means the property of yielding empirically confirmed predictions (of interest), we appear to be left not with explicit methodological rules for making veridical judgements about a theory overall, but recommendations about psychological states (confidence, acceptability) related to the usefulness of a theory. This is less than what would normally be required for a realist account of theory evaluation.

And when it comes to ‘core’ assumptions of a theory (rather than a theory overall), as Mäki himself says: ‘F53 admits that many of the central assumptions of the economic theories challenged by their critics are false’ (p. 101, emphasis added). F53 gives a number of examples of theoretical devices that are not just idealisations or judicious abstractions of aspects of reality, but rather are outright inventions: one says that it is as if entrepreneurs perfectly calculate maximum returns because, in assuming that they really do, one would be making a self-evidently false claim; one says that it is as if expert billiard players calculate their shots by consciously using Newtonian theorems because, in assuming that they do, one would be making a self-evidently false claim; one says that it is as if leaves on a tree calculate positions that maximise their exposure to sunlight (and move accordingly) because, in assuming that they do, one would be making a self-evidently false claim. Now, these examples involve a good deal of fictionalising and abstracting from the details of reality. They are very ‘simple’ (and thus ‘virtuous’) in F53’s terms because one requires only a trivial amount of prior empirical knowledge – indeed one needs to know little more (empirically) than firms produce and sell commodities to make profit, that the distribution of leaves on trees bear a certain empirical relation to the direction of sunlight, and that expert billiard players usually sink billiard balls.2121. Again, this does not necessarily imply that one requires little knowledge per se. The fictional mechanism which does the work of linking the prior empirical knowledge to a prediction may indeed require a substantial amount of non-empirical knowledge. In each of F53’s examples, one may require, say, knowledge of Newton’s laws of motion, Euclidean geometry and the calculus for optimisation under constraints. What one apparently does not need much of is detailed knowledge of, say, actual cognitive processes, or actual biological processes. It is also possible that knowledge of pre-existing synthetic presuppositions determines whether an ‘as if’ formulation is deemed appropriate. As Lehtinen (2013 Lehtinen, A. (2013). Three kinds of ‘as-if’ claims. Journal of Economic Methodology, 20, 184–205.10.1080/1350178X.2013.801560
[Taylor & Francis Online], [Google Scholar]
, p. 10) points out, ‘One does not say, for example, that “stones fall as if swallows do not migrate to warmer territories during the autumn”, because one presupposes that the migration behaviour of these birds could not affect the stones in any way’. However, a presupposition or ‘commonplace’ does not necessarily count as empirical knowledge of the world. It may simply be a convention based on a pervasive ideology. For some people it is permissible to say that trees sway as if prostrating to God; for others, it is an impermissible formulation. In any case, for F53, pre-test presuppositions need not direct theorisation. Say an intrepid mathematician constructs an equation which aims at predicting the terminal velocity of falling stones using flight patterns of a class of birds when feeding. Despite the fact that it would (one assumes) violate our presuppositions about causal relations, F53 would not forbid one from saying: ‘stones S fall as if they were hungry birds B’. Furthermore, the formulation would be deemed acceptable (from F53’s perspective) if it were subsequently validated.
View all notes
What one apparently does not need to know is the actual causal mechanisms which make them behave in the ways that they do. One would certainly be hard pressed to seriously entertain the possibility of classifying the claims in these examples as idealisations that refer to (let alone represent) underlying essential causal mechanisms. These examples seem to be obvious cases of out-and-out ‘as if’ fictions. Indeed, the examples are so manifestly absurd (in realist terms) that one gains the distinct impression that F53 is sanctioning the use of whatever ‘as if’ fictions one wishes, as long as they enable one to make ‘sufficiently accurate predictions’ – fictions, incidentally, that will be classified as ‘sufficiently good approximations’ for precisely this reason (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 15).2222. It is true that the passages in F53 citing the example of Galileo’s law of fall entailing an ‘as if’ assumption about vacuums can easily be interpreted in a realist manner as a negligibility assumption. But as Mäki himself points out, when it comes to making the analogical inference to profit maximising entrepreneurs, F53 fudges the analogy and opens the backdoor to instrumentalism (cf. Reiss, 2010 Reiss, J. (2010). The methodology of positive economics: Reflections on the Milton Friedman legacy. Erasmus Journal for Philosophy and Economics, 3, 103–110.
[Crossref], [Google Scholar]
).
View all notes

It will not do, as Mäki does, to ignore these examples just because they contradict the realist presuppositions of the reader-cum-rewriter. These examples should not be overlooked because the very purpose of the examples is to clarify the meaning of F53’s general methodological statements that, by virtue of their generality, could be misinterpreted. By dismissing the illustrations, Mäki robs himself of an important textual means of correctly interpreting F53’s general statements.

That said, the ‘general passage’ Mäki cites on page 40 of F53 about the nature of science per se does have a distinctly realist ‘ring’ to it. To repeat it:

A meaningful scientific hypothesis or theory typically asserts that certain forces are, and others are not, important in understanding a particular class of phenomena. It is frequently convenient to present such a hypothesis by stating that the phenomena it is desired to predict behave in the world of observation as if they occurred in a hypothetical and highly simplified world containing only the forces that the hypothesis asserts to be important. (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 40)

A legitimate question to ask, however, is whether the passage must have that realist ‘ring’. Is its meaning really transparent?

The crucial phrase in the passage is ‘certain forces are, and others are not, important in understanding’. What is the precise meaning of the key terms in this phase? One may be tempted to elaborate on ‘forces’ as metaphysical causal forces or causal powers. This is a natural intuition these days, thanks partly to the return of robust defences of scientific realism in the philosophy of science. But we must be cautious because the precise meaning of the term is not laid bare anywhere in the text – a text that was written at a time when such realism was near-invisible. Given that F53 cites Alfred Marshall’s work as an instructive inspiration (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
; pp. 7, 34–38, 41), we might draw a line back to the master’s use of the term in order to ascribe his meaning to F53’s use of the term. The problem with this strategy for Mäki’s approach, however, is that we are not supposed to appeal to the minds of authors (let alone secondary authors) to impute the text’s meanings. But even if we did, there are still problems. First, Marshall (1920 Marshall, A. (1920). Principles of economics (8th ed.). London: Macmillan.
[Google Scholar]
) sometimes used the term to describe a psychological concept, but according to Hirsch and de Marchi (1990 Hirsch, A., & de Marchi, N. (1990). Milton Friedman: Economics in theory and practice. Ann Arbor, MI: University of Michigan Press.10.3998/mpub.13240
[Crossref], [Google Scholar]
), Friedman almost always ruled psychological explanantia out-of-bounds. Second, when he was referring to market equilibrium, Marshall seems to have drawn an analogy to ‘force’ in the physical sciences. But, in this case, we may still wonder whether he was appealing to a realist or an instrumentalist sense of the term because the concept of ‘force’ in physics has fluctuated between realist and instrumentalist interpretations; in fact, a strong instrumentalist sense of ‘force’ – that it was actually a mental or theoretical construct – seems to have been dominant at the time F53 was written (Coelho, 2010 Coelho, R. L. (2010). On the concept of force: How understanding its history can improve physics teaching. Science & Education, 19, 91–113.10.1007/s11191-008-9183-1
[Crossref], [Web of Science ®], [Google Scholar]
). Turning to the term ‘understanding’, the waters are just as muddy. This term has a number of different meanings,2323. The Shorter Oxford English Dictionary offers the following: comprehending or apprehending the meaning or importance of something; comprehending or realising a fact; knowing by information received; accepting as true or existent; accepting or believing as a fact without knowledge; believing or assuming on account of information received or by inference; interpreting or viewing something in a certain way.
View all notes
and F53 does not specify which one it is using. It could mean that ‘forces’ help one, for example, to comprehend how a phenomenon comes about; or to accept that a phenomenon is of a certain kind; or to interpret a phenomenon as a genuine pattern and not a statistical artefact. Each of these can, with sufficient ingenuity, be given a realist or a non-realist gloss. This is evident from the fact that there are instrumentalist and logical empiricist uses of the term (e.g. Toulmin, 1953 Toulmin, S. (1953). The philosophy of science: An introduction. London: Hutchinson's University Library.
[Google Scholar]
, p. 165; Humphreys, 2000 Humphreys, P. W. (2000). Analytic and synthetic understanding. In J. H. Fetzer (Ed.), Science, explanation, and rationality: Aspects of the philosophy of Carl G. Hempel (pp. 267–286). Oxford: Oxford University Press.
[Google Scholar]
). It is thus at least conceivable that the above passage, which to us may have a realist ‘ring’ to it, could be given an instrumentalist meaning. It could, for example, be taken to mean ‘certain imaginary theoretical constructs are, and others are not, important in interpreting a particular class of phenomena as exhibiting a certain pattern’. I am not suggesting that this is what the above passage does in fact mean. The point is that it is at least possible to conceive of it in an instrumentalist way because the key terms in the passage can be given instrumentalist meanings. In sum, the general passage in F53 about scientific theories, which Mäki takes to be clearly supportive of a realist interpretation, is not quite as decisive as it may first appear.
 楼主| 发表于 2017-9-12 22:37:16 | 显示全部楼层
3.5. Ontological and epistemological commitments

Perhaps the strongest support for a realist interpretation of F53 comes from the passage which says that science seeks to ‘reveal superficially disconnected and diverse phenomena to be manifestations of a more fundamental and relatively simple structure’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 33). Mäki glosses this as ‘scientific theories are required to capture those fundamental structures [of reality]’ (p. 106) such that disparate phenomena are ultimately ‘unified’ by a single causal explanation.

When it comes to F53’s actual advice however – returning to the points made in 3.3 – we find that F53 does not say judgements about the adequacy of a theory are to be based on whether they match an underlying reality, let alone unifying causal mechanisms. Rather, F53 tells us that such judgements should be based on whether theories are ‘sufficiently good approximations for the purpose in hand’. And again, what counts as ‘sufficiently good approximations for the purpose in hand’ is determined exclusively by using a simple theory to derive ‘sufficiently accurate predictions’ over a wide range of phenomena of interest to us – the wider the range and the simpler the theory, the better. With this in mind, we can make better sense of F53’s reference to the ‘essential features of complex reality’. By the process of testing that F53 advocates, certain ‘features’ come to be (fallibly) judged as ‘essential’ to the prediction of phenomena of interest to us. In this sense, these features of reality are not metaphysically essential-in-themselves, but rather are instrumentally essential-for-us – that is, they are essential for our particular purposes at a given point in time. With a change in purpose, despite there being no change in reality per se, there would be a change in what counts as ‘essential features of reality’. This being so, the unification of disparate observable phenomena which Mäki speaks of would instead be glossed as true derived implications over the maximum range of phenomena of interest to us from a maximally parsimonious theory.2424. It is worth recalling that the ‘Logical completeness and consistency’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 10) of theories and their implications is referred to immediately after F53 defines ‘fruitfulness’. It should also be recalled that with respect to testing the worth of a theory, F53 tells us that ‘further testing [of a theory] involves deducing from it new facts capable of being observed but not previously known and checking these deduced facts against additional empirical evidence’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 13, emphasis added). Clearly, the implications of a theory – its predictions – are logical derivations.
View all notes
But that being the case, Mäki’s explanatory unification in fact reduces to ‘derivational unification’, which Mäki suggests coheres best with a non-realist stance.

Indeed, Mäki admits that F53’s example of conscious leaves does not conform to the notion of a unified causal explanation: ‘realist hopes are undermined as no role is supposed to be played by what is justifiably believed to be the case in the real-world causation of leaves’ behaviour’ (p. 107). In this case, the presumption of an underlying unity to reality per se is not decisive to theory choice (or a motive for theory construction). Rather, what is decisive is ‘the range of conclusions that can be derived … and the range of theoretic connections they exhibit’ (p. 107). As Mäki says, this is not explanatory unification; it is rather derivational unification.

Finally, there is a simple textual problem with Mäki’s suggestion that F53 advocates unifying explanations over derivational ones. Mäki says that F53’s use of the word ‘fruitfulness’ is synonymous with explanatory unifying power (p.109). Unfortunately, this simply does not match F53’s actual definition of this term. According to F53, a theory ‘is more ‘fruitful’ the more precise the resulting prediction, the wider the area within which the theory yields predictions, and the more additional lines for further research it suggests’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 10). Fruitfulness is a criterion that makes the precision and scope of the deduced implications of a (simple) theory a decisive virtue. It is not even about correct prediction, let alone the expression of some causal mechanism that may be ‘behind’ the observed phenomena; it is almost entirely a matter of extending deductions (predictions) into new domains using the same basic theoretical apparatus. Thus, what Mäki takes to be the very term for explanatory unification in F53 actually turns out to be a prescription for promissory derivational unification par excellence.
 楼主| 发表于 2017-9-12 22:37:30 | 显示全部楼层
4. Conclusion

Mäki’s rereading-rewriting of F53 crucially depends on a conscious selection of passages from the text. This is the Achilles’ Heel of his interpretation because it throws into relief the fact that the selection process is governed by the very realist position Mäki seeks to ‘find’ in the text. It will not do to counter that at least the passages selected are the ‘central’, ‘key’ or ‘essential’ ones because that claim itself is also informed by the presupposed realist position. Further, even accepting that some of the passages selected are ‘central’, since most of them are ambiguous, in order to ‘infer’ realist meanings one must, again, presuppose an underlying realist position. In short, at every turn, Mäki’s interpretative method renders his rereading-rewriting of F53 vulnerable to the charge of circularity. This is not to say rival exegetical efforts do not involve a similar circularity. It is just to say that it is more obvious in Mäki’s case because he is candid about his selectivity.

Space constraints forbid an elaboration and defence of what might be hubristically called a more ‘accurate’ interpretation of F53. Perhaps it is permissible however to conclude by briefly saying something both programmatic and polemical on this topic. In his watershed discussion of F53, Boland (1979 Boland, L. A. (1979). A critique of Friedman's critics. Journal of Economic Literature, 17, 503–522.
[Web of Science ®], [Google Scholar]
, p. 509) does not just say that F53 expresses an instrumentalist philosophy of science. He says that F53 is ‘an instrumentalist’s argument for instrumentalism’; that is, F53 implicitly deploys an approach to methodology itself that is instrumentalist (call it ‘meta-instrumentalism’), wherein each of F53’s arguments on their own, whatever their philosophical origins or implications, are deemed useful (sufficiently good) for the overarching ‘purpose in hand’. What was that overarching purpose? As has been argued by others, it seems to have been to defend, in the service of certain disciplinary, political and ideological ends, a Marshallian-inspired theoretical apparatus by undermining scientific status of rival theoretical and methodological projects (cf. Katouzian, 1980 Katouzian, H. (1980). Ideology and method in economics. London: Macmillan.10.1007/978-1-349-16256-7
[Crossref], [Google Scholar]
; Mulberg, 1995 Mulberg, J. (1995). Social limits to economic theory. London: Routledge.
[Google Scholar]
; Freedman, 2008 Freedman, C. (2008). Chicago fundamentalism: Ideology and methodology in economics. Singapore: World Scientific.10.1142/6809
[Crossref], [Google Scholar]
; Backhouse, 2009 Backhouse, R. E. (2009). Friedman’s 1953 essay and the marginalist controversy. In U. Mäki (Ed.), The methodology of positive economics: Reflections on the Milton Friedman legacy (pp. 217–240). Cambridge: Cambridge University Press.10.1017/CBO9780511581427
[Crossref], [Google Scholar]
). As such, Friedman was in effect, agnostic vis-à-vis different arcane ontological and epistemological positions. Such a meta-instrumentalist approach does not seek to prohibit an economist from holding, say, realist beliefs about F53’s nominated theories. Nor does it prohibit methodologists from giving realist or pragmatist or falsificationist or positivist interpretations to various passages in F53 itself. Nor does such an approach require strict adherence to just one philosophical framework. In short, F53 may well be interpreted as expressing, say, a useful realist position here or a useful non-realist position there – and in this respect, it is an expression of a meta-instrumentalist approach to methodology.

The philosophically insubstantial nature of this meta-instrumentalism helps us account for the fact that it is so difficult to capture the ghostly ‘reality’ of what the author is ‘actually saying’ in his text. Friedman – by his own admission, and in a characteristically meta-instrumentalist way – was quite happy for people to believe it is ‘entirely correct’ to call F53 instrumentalist, or that it is ‘almost identical’ to John Dewey’s pragmatism, or that it is ‘entirely consistent’ with falsificationism. Indeed, this opportunistic strategy is captured in Friedman’s final word on the entire matter: ‘I feel like a proud father who has a large brood of bright children – all of them right, all of them wrong, and all entitled to his or her own views’ (Friedman, 2009 Friedman, M. (2009). Final word. In U. Mäki (Ed.), The methodology of positive economics: Reflections on the Milton Friedman legacy (p. 355). Cambridge: Cambridge University Press.10.1017/CBO9780511581427
[Crossref], [Google Scholar]
). All the interpretations are acceptable as long as they serve the purpose of helping to persuade the discipline to utilise Friedman’s nominated theoretical apparatus, with the same ‘sufficiently good’ ‘tight priors’ conducive to a certain kind of ideologically informed policy consensus (cf. Reder, 1982 Reder, M. W. (1982). Chicago economics: Permanence and change. Journal of Economic Literature, 20, 1–38.
[Web of Science ®], [Google Scholar]
). That is, the text is best seen as a tool or weapon designed to direct the theoretical trajectory of economics, and so contribute to the grander goal of shaping a disciplinary consensus over ‘economic policy in the Western world’ (Friedman, 1953 Friedman, M. (1953). Essays in positive economics. Chicago, IL: University of Chicago Press.
[Google Scholar]
, p. 6). It is perhaps the neglect by methodologists of this overarching purpose that has led to the fifty-odd year search for ‘the truth’ in F53 rather than the truth about it.
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

关于我们|Archiver|手机版|小黑屋|人大经济论坛 ( 京ICP备05066828号-20 京公网安备 11040202430141号 )

GMT+8, 2019-10-24 00:53 , Processed in 0.042218 second(s), 23 queries .

Powered by Discuz! X3.4

© 2001-2017 Comsenz Inc.

快速回复 返回顶部 返回列表