I got a degree in 'Analytical & Descriptive Econ' at the LSE in 1982. It was explained to me that Descriptive Econ was like Accountancy. You have to give a true and fair view based on pre-specified objective criteria. If you are in Advertising or Marketing, you can select things to highlight in a subjective manner. But you can't do that as a Descriptive Economist or Accountant.
Police officers are trained in observation. They are required to give a description of suspects. They do so by mentioning height, weight, build, colour of eyes etc. They are discouraged from saying things like 'what I choose to remember about the suspect was his air of elegant weariness. More than that, I can not say.'
Analytical Econ, by contrast was about creating a Structural Causal Model and seeing how well it fits a particular data set. Again, if your purpose is alethic rather than polemical, there can be no question of subjective choice or whim or fancy regarding what observations you select and what you discard.
A description of a sick man- or a society suffering some calamity- may move us to pity or anger at injustice. A completely fabricated and false description may be more effective, as propaganda, than alethic description. But stupid lies aren't 'descriptions'. They are crazy fantasies which only a fool would give credence to.
On the basis of a description, a diagnosis can be made and a prescription applied. If the prescription succeeds completely, it may be that only one fact in the description was relevant. A description which omitted it would be useless. But, a description which hightlighs it and omits other things might still be acceptable.
Positive Economics is Analytical. Descriptions of an economic situation may lead to the adpotion of a Structural Causal Model capable of diagnosing, prescribing or making testable predictions. However, the Model may not itself utilize or contain any information arising from the description.
Consider the Marxist description of the Economy as involving Employers creating profit or surplus value by forcing workers to put in longer hours. It is false. However, it is true that 'capacity utilization' is linked to profit. But this 'analysis' (based on a false description) is highly mischievous. During the Great Depression, it would say 'reduce the power of Trade Unions so that money wages can fall and increased surplus value can be extracted.' But this would also be the recommendation of the believer's in Say's law! They would say, prices have to fall to compensate for the exogenous money supply shock. Trade Unions are evil because they create 'downward stickiness of wages'.
Descriptively, this may have some truth though there is a wider problem of administered pricing, however, analytically, the thing was garbage because lowering all prices (including re-writing all rental agreements etc) wasn't feasible within a reasonable time-scale.
Once re-armament started,there was increased capacity utilization but it went hand in hand with either 'repressed inflation' (i.e. longer queues & smaller rations) or hyperinflation unless a capital levy was politically feasible. Interestingly, two types of 'description'- one Marxian (about pressure to work longer hours) and one Keynesian (about aggregate demand & capacity utilization)- could happily coexist. This was because employers genuinely want to get more work out of those they pay (even if they are charitable institutions). Moreover, in making capacity utilitzation decisions, enterprises genuinely do look at aggregate demand.
However, the analytical theory that sat on top of this dual descriptive econ broke down once the Phillips Curve started to shift or break down altogether. This was no great scandal for Econ because better descriptions supporting better analysis was possible through 'micro-foundations of Macro'. On the other hand, increased financialization meant that smart Economists got rich in the City. The Cloister contained the cretinous.
Since very stupid and useless people like scolding, Normative Economics- which involves condemning or praising states of the world- got a boost around this time.
Since scolding, ideally, applies to actual, observable behaviour, Normative Econ is linked to description in a direct manner which is not the case with Positive Economics. Sadly, no alethic descriptive econ supports any normative analysis. Thus paranoid fantasy has replaced 'descriptive economics'.
Still, speaking generally, a normative approach would prefer to rely on alethic, or plausibly alethic, descriptive statements- e.g. Capitalism caused 10.3 billion workers to starve to death as is recorded in the Census- rather than obviously crazy lies- e.g. Capitalism caused 10.3 billion workers to be sodomized to death by my Mum. Fuck you Mummy! Fuck you very much!
However, this is not a really important point. Normative Economists are perfectly comfortable repeating any type of lies. Since their purpose is to scold, they don't care that their work has no predictive power. It must be said, this is not a choice on their part. It is a pathology. They are stupid and useless & thus their purpose in life is to make a nuisance of themselves.
In a 1980 paper, Amartya Sen wrote of 'Description as Choice'
Like much of his work, it was based on the logical fallacy of treating a predicate (a property, attribute, or description) as if it were a concrete object or an independent entity. This is known as the fallacy of reification (also called hypostatization, the fallacy of misplaced concreteness, or concretism).
To compound the fallacy, Sen continually uses the word 'description' to refer to things which aren't descriptions at all.
In other words, his essay is vitiated by logical errors & sheer ignorance or wilful defiance of the rules of the English language.
Bertrand Russell distinguished between 'Knowledge by Acquaintance' - direct, sensory awareness of things- and 'Knowledge by Description'- indirect knowledge gained through conceptual descriptions.
Where knowledge is direct or immediate, scope for choice exists. Where it is indirect or 'by Description', such is not the case. We may say that choices are made while describing a thing. But unless those choices render the description useless for conveying knowledge, no great scope for choice exists.
A quite separate point has to do with the Watanabe ugly duckling theorem which, in 1969, showed that statistical classification can't be free of bias. But this does not mean any honest, alethic, work in Statistics is based on giving free rein to subjective choice.
The truth is no description, properly so called, is based on Choice. We may say there is a bias- but that is a pathology just like having cancer. No one chooses to have it. Moreover, it may be amenable to amelioration.
DESCRIPTION is sometimes contrasted with prescription,
The description of a disease is different from the prescription for it. I can give a good description of my symptoms to my Doctor. I can't give myself the same prescription as he gives me because I lack medical training.
and at other times with prediction.
A prediction may be based on a description. But they are entirely different things. Because of 'multiple realisability' the prediction may be the same for things which have very different descriptions.
While the former differentiates descriptive statements from value judgments and imperatives,
It does not do so. A prescription is not a value judgment if it is protocol bound & based on purely alethic propositions. Thus a Doctor who prescribes the correct medicine to me is not making the value-judgment that I should continue to live. He is merely doing his duty which has been specified in wholly positive, not normative, terms.
the latter differentiates it from predictive statements.
There is no need to do so. If two people jump off a cliff, we predict both will die even if their physical description are very different
Philosophical discussions in the social sciences have tended to concentrate on prescriptive and predictive exercises, and as a consequence the methodological issues involved in description have remained largely unexplored.
Not all philosophical discussions are foolish. Social Sciences make progress when their methodology improves such that better observations are made possible and thus better descriptions become available. Sometimes such methodological improvements involve 'open questions' in Math and thus the subject is intrinsically philosophical (till some advance in STEM subjects closes the question).
The object of this paper is to take up some of these issues. Boundaries are not always clear-cut between prescription and description,
They always are. No doubt, one may disguise the one as the other so as to give or avoid offence. Thus I I would say to Prof. Sen 'you are looking peaky. Do dine heartily on dog turds tonight so as to return to your usual ruddy good health.' This is neither a description nor a prescription. It is an insult. The suggestion is that Sen eagerly devours what emerges from the backsides of bow-wows.
... I believe that boundary questions are sometimes taken to be more important than they are.
Observing boundaries lessens pointless conflict. That's important.
Intellectual interest in these issues may distract attention from the fact that imprecision of boundaries can still leave vast regions without ambiguity.
That is wholly irrelevant.
It is indeed possible to say a good deal about China and India without asserting that there are no ambiguities as to where the boundary between the two countries lies.
Yet, it is only because there is an ambiguity regarding boundaries that both countries have squandered a lot of blood and treasure. The 1962 war had a disastrous impact on the Indian 5 year plan- as Sen knows very well.
It is fair to say that description as an intellectual activity is typically not regarded as very challenging. To characterize a work in the social sciences as “purely descriptive’ would not normally be regarded as high praise.
Whereas Sen-tentious shite received high praise because he was a brown monkey.
The reason for this—at least in part—must rest in the idea that description is largely a matter of mere observation and reporting, or reading other people’s reports and summarising—at best systematizing. Whether a descriptive statement is acceptable could be thought to be dependent on its correctness, and that could be resolved simply by observing. I would like to begin with what I think is a moderate claim, reserving the right to be controversial later on. Description isn’t just observing and reporting; it involves the exercise—possibly difficult—of selection.
Not to the diligent describer. They mention the thing because they believe it has objective salience. They may be wrong.
For example, in judging F. M. Eden’s 1797 study The State of the Poor, or Frederich Engel’s The Condition of the Working-Class in England in 1844, or John and Barbara Hammonds’ The Village Labourer 1760-1832, a good deal more is involved than just checking the truth of the individual facts recorded.
That is all that is involved. Liking of disliking a thing is a matter quite separate from judging a descriptive work on the basis of its degree of accuracy and completeness.
In fact, description can be characterized as choosing from the set of possibly true statements a subset on grounds of their relevance.
It is possibly true that Sen looks peaky. It is also possibly true that he likes eating dog turds. But my statement that Sen is looking peaky because he hasn't eaten enough dog turds is not descriptive. It is insulting. It is defamatory. Why?
'Possibly true' isn't enough. Verification of truth is required though one can specify that one's description is partial.
Moreover, a description can be given by a policeman or Auditor in which no personal discretion or choice was involved in the gaining or information under pre-specified rubrics.
Truth is—at best—a necessary but not sufficient condition for good description.
In
philosophical logic, a slingshot argument is one of a group of
arguments claiming to show that all
true sentences stand for the same thing. We can't be sure there are any necessary or sufficient conditions on any true sentence (because it may be the same as all other true sentences).
It is perhaps not an exaggeration to say that any conscious act of description contains some theory—usually implicit—about the relative importance of the various statements dealing with the subject matter. I shall call this the “choice basis of description’’.
One could equally say that one has no choice regarding one's biases. They are the cross you must bear.
I might add in passing that recent developments in formal choice theory can be fruitfully used for studying the regularities of description, but I shall not pursue this suggestion in this lecture.’
Formal choice theory was nonsense. Either there is a choice sequence of a Brouwerian type or there is no actual choice.
3 A description can be accurate without being a good description.
No.
It could be unhelpful, even useless.
In which case it is not a description at all.
We question the expert on the level of factory wages in India.
This is described by a statistical curve.
He answers: “Oh, it varies from place to place.’
He has not given you the description you asked for.
True enough, of course. We ask for more description, demanding precision. The expert now goes into details. ‘““The integer approximation of the national average wage in rupees,” he says, “‘is a prime number’’.
It isn't. That's just a lie.
I won’t belabour the point further.
You haven't made any point whatsoever.
Your child asks you: “How large is China?’’, and you reply:
Geographically, it is three times the size of India and about the same size as the US.
‘““Very large—it has 900 million people”.
Your child didn't ask you the size of China's population.
Insofar as truth is a necessary condition of good description,
Accountants speak of 'materiality'. You can ignore stuff which doesn't materially alter the picture. Sen wouldn't have been such a fool if he had studied a bit of Accountancy.
the concept of truth must be a broad one.
No. It must be utilitarian.
The second problem is, I think, a good deal more complex. Description may have objectives the pursuit of which can be helped by departures from truth—even in the broad sense.
No. Models can ignore aspects of reality, or even falsify them. What matters is whether they produce good predictions.
Milton Friedman (1953), in particular, has argued powerfully in favour of departing from truth in describing reality in the context of economic models, judging the merits of the assumptions not in terms of accuracy but in the light of their predictive usefulness.
This is Instrumentalism. Forecasting is separate from Analytical & Descriptive Econ.
I shall have something to say presently on Friedman’s methodological observations, and on Paul Samuelson’s critique of Friedman and the “F-twist’”
i.e. Friedman saying Positivism is about predictions not verisimiltude.
but at this stage I would simply note that Friedman has drawn our attention to the important possibility that an unrealistic assumption giving an inaccurate description may nevertheless be a useful one for the purpose for which the description is intended.
No. Description is completely separate. I predict that Miss Smith will die because she is on a plane which is currently crashing into the Arctic Ocean. You ask me if Miss Smith is a blonde or brunette. I don't know. I haven't seen her. All I know is that she is on the doomed plane.
Friedman was saying 'inflation is always and everywhere a monetary phenomenon'. Others were saying 'but my country is structurally different from any country you have studied! How can the outcome be the same?' Friedman would reply, 'for the same reason that Miss Smith will die whether on not she is a blonde.'
Friedman restricts his attention to predictive usefulness only, but the point can be generalized, and objectives other than predictive success might also lead to a case for departing from the truth. Examples can be constructed such that truth of descriptions conflicts with usefulness for prescriptive exercises, or even for communication.
e.g. my description of Sen as looking peaky because he hasn't noshed on dog turds.
4 A distinction must, however, be made
Sen has never made a useful distinction iin his life
between a description of something being good in the sense of being a good one to give and it being a good description of that thing.
A 'good description' means a 'true & fair' account of the thing.
The aspiring murderer demands from you a description of where his would-be victim has gone,
He is asking for information, not a description. Indeed, he may have to supply you with the description unless you too are already acquainted with the victim.
and as you point at the wrong road he proceeds in that direction with a roar.
You didn't give a description. You gave false information.
That description of where the would-be victim has gone is, I would agree, a good one to give,
Nobody has given any fucking description.
but it can hardly be accepted as a good description of where the would-be victim has gone. The goodness of a description can be judged in terms of many alternative criteria,
No. It can only be judged in one way. It is a different matter that you may prefer to act capriciously rather than to judge the thing. Such behaviour can disqualify you from doing so.
but in calling a description a good description of something, truth—in the broad sense—would seem to be a necessary condition.
Yet, such isn't the case. With respect to any type of descriptive work, there are experts in the field who judge what is good and is bad. If we don't have that expertise, we can't pronounce on the matter.
There isn’t anything very puzzling here, even though the distinction has been frequently overlooked. A good description of something reflects reality about that thing is some straightforward sense, rather than distorting it.
I describe a good Amartya Sen who eats dog turds in a good manner. To my mind, this reflects reality. However, I am not an expert on Sen's diet. It is possible that when Amartya Sen is good, he eats mice. When he is naughty, he feasts on camel droppings.
I turn now to the methodological dispute between Friedman and Samuelson.* Friedman argued that the appropriateness of an assumption must be judged not by its realism but by its predictive usefulness.
He was saying nothing about description
Furthermore, Friedman saw an inverse relationship between realism of assumptions and its predictive contribution.
No. He said more significant theories may have less realistic assumptions. But an assumption is not a description. It is either itself a prediction or it a step towards getting a prediction.
Insofar as a theory can be said to have ‘assumptions’ at all,
In Econ, every theory comes with a list of assumption
and insofar as their ‘realism’ can be judged independently of the validity of predictions,
they aren't meant to be realistic.
the relation between the significance of a theory and the ‘realism’ of its assumptions is almost the opposite of that suggested by the view under criticism.
Sadly, Econ theories aren't significant at all. They are just shite you teach. Being able to correctly predict things, on the other hand, can make you very rich.
Truly important and significant hypotheses will be found to have ‘assumptions’ that are widely inaccurate descriptive representations of reality, and, in general, the more significant the theory, the more unrealistic the assumptions (in this sense).
An assumption is not a description and vice versa. Consider the assumption of perfect competition. We can't describe a world where it obtains. Indeed, such a world is 'incompossible'. It can't obey the laws of physics which obtain in the real world.
Characterizing the approach as “F-twist’”, Samuelson has attacked Friedman’s position sharply. “If the abstract models contain empirical falsities, we must jettison the models, not gloss over their inadequacies.”
& say goodbye to Physics, Chemistry, Biology- any nomothetic discipline.
“The empirical harm done by the F-twist is this. In practice it leads to Humpty-Dumptiness. Lewis Carroll had Humpty-Dumpty use words any way he wanted to. I had in mind something different: Humpty-Dumpty uses the F-twist to say, “What I choose to call the admissable amount of unrealism and empirical invalidity is the tolerable amount of unrealism’”’.’
But if Humpty gives better predictions than any one else, we don't fucking care.
I am not presently concerned with whether Friedman is correct in thinking that an inverse relationship tends to exist between accurate assumptions and predictive success.
He didn't think that. He merely said that if a model has better predictions, and you gain by having better predictions, then you stick with it even if you find its assumptions unrealistic.
The important point to note is that if such an inverse relationship were to obtain, Friedman would be quite willing to sacrifice accuracy.
Accuracy is not available. It is simply a fact that if what matters is prediction, accurate prediction is all that matters. If description is what matters, accurate description is all that matters. They are two separate things.
Stanley Wong (1973) has pointed out that Samuelson’s critique takes insufficient note of the “instrumental” nature of Friedman’s methodological position, and the latter’s overriding concern with prediction. This, I believe, is indeed the case, but the two views of description differ on an issue that goes well beyond the question of the predictive focus.
There is only one view of description. Sen is making this shit up.
No matter what the aim happens to be, as long as it is not—or does not include—making only accurate statements about observed facts, a conflict can conceivably arise between descriptive accuracy and the aim of the description.
No. Descriptive accuracy is objective. It is wholly independent of the subjective aim.
The issue is a good deal more general than the so-called ‘“‘F-twist”, and concerns the conflict between aims of description and the truth of statements that can be used to further these aims.
Description is a word. It is not a being with aims or dreams or preferences or a philosophy of life. Some descriptions have verisimilitude. If so, we say they are good. This does not mean that they regularly condemn Narendra Modi. It doesn't even mean that they offer gratuitous rape counselling to senior citizens. It just means that we are pleased with them.
Another different motivation is involved in Nicholas Kaldor’s’ advocacy of departure from truthful description, in favour of what he calls “stylized facts”.
i.e. assumptions re. what would, approximately, be the case, ceteris paribus.
These aren’t facts in the sense of being true, but they communicate rapidly certain things that are close enough to being true, but avoid a plethora of details.
that is what an approximation is.
An example: that in the Western economies the capital output ratio remained stationary over many decades and so did the share of wages in national income.
This is clearly false because of the business cycle.
Not exactly true for all the alluded economies. In fact, not exactly true for any of the alluded economies. Still, if the variations are relatively small in comparison with other, related variables, the simplification will have certain obvious advantages.
So will lying.
I have to report that in this case the advantages were probably rather unequally shared, with the biggest share going to Kaldor himself, who used these stylized facts to illustrate his well-known models of growth and distribution. But that wasn’t all there was to it, and some of the stylized facts undoubtedly succeeded in enhancing comprehension, drawing attention to important observed regularities, even though they were not quite accurate. However, the important issue here isn’t whether these particular simplifications were all informative without being misleading but that the method of using “‘stylized facts’ can indeed have this property.
So can lying or spinning a yarn or telling a 'just-so' story.
There is nothing exceptional to economics in making stylized facts a possible and potentially efficient method of understanding reality.
or lying that you do.
Fiction is a general method of coming to grip with facts.
Only in the sense that lying is a general method of revealing the truth.
There is nothing illegitimate in being helped by War and Peace to an understanding of the Napoleonic wars in Russia, or by Grapes of Wrath to digesting aspects of the Depression.
Nor is there anything legitimate about it. Tolstoy was completely wrong about Napoleon & Kutuzov. There is no aspect of the Depression illumined by Steinbeck. The fucker never even visited Oklahoma. Also American women didn't breastfeed starving strangers.
There is no reason why descriptive statements in economics have to aspire after mechanical accuracy even when it conflicts with comprehension and absorption.
Descriptive Econ or Accountancy must be accurate. Otherwise it gets the fucking sack. What Sen did wasn't descriptive. It was masturbatory or time wasting.
There is, of course, an obvious objection to presenting non-facts dressed up as facts, but there is no need to do this once non-facts are accepted as ligitimate descriptive instruments themselves.
Once you accept lies as truth, you don't need to bother with truth. Also, by accepting dog turds as tasty food, you can subsist wholly on dog turds.
Such a description of something will be good in the sense of being useful, but—as already explained—must not be confused with its being a good—or realistic—description of that thing.
I describe dog turds as tasty food. Sen is obliged to eat them.
Samuelson’s argument that ‘we must jettison the models” that “contain empirical falsities’’ seems to be based on
stupidity. All models in the natural sciences 'contain empirical falsities'.
a very narrow view of objectives of description. However, Samuelson wasn’t only objecting to the use of empirical falsities in models, but also disputing Friedman’s use of words.
No. He was saying that Friedman is making up his own criteria of success. This wasn't true. The criteria was that predictions got verified by outcomes.
The reference to Humpty-Dumptiness deals with this. This relates closely to our distinction between a description that is good and a good description of something.
They are one and the same. The meaning is 'this description has high verisimilitude' or what Popper called 'truth-likeness'.
Friedman not only defends the use of inaccurate assumption in predictive theory, but also employs a new concept of realism. It is “‘the test by prediction” in classifying alternative assumptions “‘as more or less realistic” that Friedman embraces and which he sees others being ‘‘necessarily driven” to accept (p. 33).
Sadly, when Friedman's predictions failed to materialize, his theory was abandoned. This what was happening around the time Sen published this.
The predictive usefulness of an inaccurate description may well justify the choice of that description, but why does it make it a more realistic description?
We chose the prediction, not the description, if we are only concerned with the outcome. If we don't care about the outcome but for some reason want a picture of how things were, we go for the description.
The “instrumental” nature of Friedman’s position may be adequate in defending the choice of such an assumption, but to call it “‘realistic” is a separate assertion, violating the important criterion that a description must be true, in the broad sense, to be realistic.
Sen assumes that it is easy to say which description is relistic. But, he himself seems to think that Tolstoy & Steinbeck gave realistic descriptions whereas, notoriously, the opposite was the case. Napoleon was a freakin' military genius. Okies weren't evicted by evil bankers. FDR's agricultural adjustment act incentivized landlords to eject tenants so as to lower their cotton acreage. Government subsidies enabled them to replace workers with tractors. About a quarter of all tenant farmers in the South West upped sticks during that decade for this reason. Thus Steinbeck was completely wrong about everything. Only a cretin like Sen would have bought into his crude propaganda.
Samuelson’s objection to this violation cannot be dismissed by a pointer to the “instrumental” character of Friedman’s economic methodology.
Sure it can. Samuelson was wrong. Friedman was right. The Phillips curve didn't really represent a trade-off. Anyway, you can could always change the assumptions to make them more appealing to any particular audience. 'Micro-foundations' was a way to reverse engineer things.
The defence that Friedman provides of this characterisation of ‘“‘realism”’ mainly takes the form of denying the viability of any alternative conception. A theory or its “assumptions” cannot possibly be thoroughly “realistic” in the immediate descriptive sense so often assigned to this term.
It must have lower Kolmogorov complexity than what it relates to.
A completely “‘realistic” theory of the wheat market would have to include not only the conditions directly underlying the supply and demand for wheat but also the kind of coins or credit instruments used to make exchanges; the personal characteristics of wheattraders such as the colour of each trader’s hair and eyes, his antecedents and education, the number of members of his family, their characteristics, antecedents, and education, etc.; the kind of soil on which the wheat was grown, its physical and chemical characteristics, the weather prevailing during the growing season; the personal characteristics of the farmers growing the wheat and of the consumers who will ultimately use it; and so on indefinitely.
No. A particular market for wheat would indeed have particular determinants of Demand, Supply, etc. However, hair colour is never relevant. The question is whether outcomes in a particular wheat market depart too much from what the existing theory, absent any local knowledge, predicts. This does not mean, for some other purpose- e.g. ecological- that interference in a particular market might not be justified.
Any attempt to move very far in achieving this kind of ‘‘realism” is certain to render a theory utterly useless.
No. But it isn't parsimonious or it needlessly adds noise to signal.
In assessing this objection it is necessary to consider the distinction between realism in the sense of ‘“‘nothing but the truth” and that in the sense of ‘the whole truth”.
They are one and the same. Commission & omission arise in testimony. But the truth is not testimony. It is a predicate applicable to testimony.
An assumption can be realistic in that it is true without the claim being made that it is exhaustive in capturing all aspects of the reality.
all relevant aspects. Truth means relevant truth.
Advocates of realism in the sense of “nothing but the truth” need not demand “the whole truth’.
There really is no difference between the two. The truth of the matter is all relevant information. Selective testimony is not true testimony. It is biased.
The dissatisfaction with Friedman’s position on the part of critics such as Samuelson does not arise from Friedman’s rejection of “the whole truth’, but from his rejection of “nothing but the truth’’.
No. He says neither matter. Only the quality of predictions is germane. Remember, it is only in a court of law- where a civil or criminal matter has to be adjudicated that testimony under oath of this type is required. But most matters are not justiciable.
The distinction is important for Friedman’s discussion of departures from realism and the criteria of acceptability of such departures. What is the criterion by which to judge whether a particular departure from realism is or is not acceptable?
Unless the matter is justiciable, there re no such criteria. Friedman was right to say that all we care about is prediction. We don't need to know how the sausage is made. It could be done within a 'black box' for all we care.
Why is it more “unrealistic” in analysing business behaviour to neglect the magnitude of businessmen’s costs than the colour of their eyes?
Because, in business 'costs' matter. Eye colour does not. Sen may suffer from some major mental malfunction such that he can't tell what is relevant and what is irrelevant for businesses. But this is a very rare malady indeed.
The obvious answer is because the first makes more difference to business behaviour than the second; but there is no way of knowing that this is so simply by observing that businessmen do have costs of different magnitudes and eyes of different colour.
Only if there is no way of knowing what knowing is or what no is or what is is. The problem with appearing a fucking cretin is that others can be more cretinous yet.
Clearly it can only be known by comparing the effect on the discrepancy between actual and predicted behaviour of taking one factor or the other into account.
No. It is known a priori and arises out of 'theory of mind'- much of which is hardwired.
Even the most extreme proponents of realistic assumptions are thus necessarily driven to reject their own criterion and to accept the test by prediction when they classify alternative assumptions as more or less realistic."
There is no such compulsion. Friedman & Samuelson were featured in Newsweek. I suppose they wrote entertainingly enough, but- by 1980- the debate was over. Samuelson was wrong about the Phillips curve. Going forward you had increasingly arcane Fintech of a type senile shitheads like Sen could not follow.
However, Friedman’s concentration on predictive success as the only objective he considers,
for Positive Economics
also makes his treatment of the choice basis of description a very limited one.
because Positive Econ has a limited aim. Sen's normative econ, wants descriptions (false or foolish ones) for a polemical purpose. Friedman is like a Doctor who wants to cure the patient. Sen is like a paranoid nutter who attributes the illness suffered by the patient to the Devil or Capitalism or the fact that Manmohan Singh is Prime Minister.
Description may well be geared to some objective other than prediction, e.g., normative analysis,
lying and scolding
or efficient communication,
in which case prediction is what matters. You use the descriptive term you predict will pick out the right object for your interlocutor. Schelling focal theory is the relevant solution concept.
or even satisfying idle curiosity. Even the preference for concentrating on businessmen’s costs rather than on the colour of their eyes—to return to Friedman’s example—may arise from some non-predictive motivation. For example, within the normative perspective, the businessmen’s costs might be relevant to some notion of desert, e.g., a person deserves to get a reward for the costs he has undertaken,
I deserve to get a reward for phoning up Dr. Sen in the middle of the night to inform him of places where he can get fresh dog turds. This is because spending my time in this manner imposes an opportunity cost on me represented by money I could have earned doing something else.
and this may give no place at all to the colour of the person’s eyes. The nature of economic questions reflects a variety of interests that the subject has to cater to. Has the distribution of income turned more unequal in country x?
Yes. Also Capitalists are invisibly sodomizing everybody. Manmohan Singh is the culprit. Fuck you Manmohan! Fuck you very much!
Are the residents of country y better fed today than before?
No! They aren't eating enough dog turds.
Is modern technology getting easily absorbed in country z?
Yes. Technology is easy to absorb- if you have a bit of cash.
Has trade unionism influenced the distribution of income under capitalism?
This went into reverse during that period.
Are there business cycles still in the Western economies?
This is about prediction and is positive
How large is the Soviet Union’s international trade?
Descriptive Accountancy
How should we plan the use of exhaustible resources?
Normative.
Does progressive taxation affect incentives?
Predictive- thus Postivie
Will the expansion of money supply—this one in honour of Milton Friedman—expand prices proportionately? ‘
Predictive- thus Positive
This bag of assorted questions reflects a plurality of motivation that is not reducible to some kind of single “ultimate objective’.
Each motivation is reducible to one and only one objective. This means a choice has to be made unless a pathology- like that of Sen's- supervenes such that only stupid lies and peurile scolding is produced.
They represent predictive investigation, normative concern as well as curiosity related to a variety of interests from sympathy to satisfying one’s spirit of inquiry.
Only if they don't have an objective. But a motive without an objective isn't a motive at all. It is merely some incohate state productive of nothing.
A monolithic characterization of the objective of description will leave us with a methodology ill-suited to the subject.
No. It will enable Econ to be useful rather than a platform for a brown monkey to gas on endlessly.
The multiplicity of motivation has been discussed so far in the particular context of departures from accuracy, but it is of course a much more central issue than this particular context might suggest. Even when no departures from truth are involved, there is—as already discussed—the problem of selection, choosing from the set of true statements a subset of relevant ones.
There is a feedback loop regulated by Occam's razor such that stuff with little predictive power gets deselected.
Criteria of relevance are crucial to this choice basis of description going well beyond the issue of truth. Methodological discussions in economics have tended to concentrate on predictive and prescriptive concerns, and seem to have had the effect of ignoring other motivations that stimulate inquiry.'
E.g the motivation of brown monkey to just gas on and on and on in a completely useless and stupid manner.
This bias has had a rather inpoverishing role on the descriptive traditions in economics. The choice basis of description may relate to the curiosity that precipitates inquiry, and this curiosity need not reflect predictive or prescriptive interests.
It may reflect the desire of a brown monkey to gas on endlessly.
Indeed, even the use of language can relate to selection based on other types of curiosity. In response to the question: “Is China a large country?”’,
the answer is 'yes. By area, it is the third or fourth largest'.
we accepted the answer: “Yes, it has 900 million people’,
This is the answer to the question 'does China have a large population?'. The answer, at that time, was 'yes. It is the most populous country in the world. '
basing the notion of the size of country on the size of its population. However, China has less surface area than Canada, fewer tigers than India, fewer polar bears than the Soviet Union, and—according to some early enthusiastic accounts—no flies at all.
All of which is irrelevant.
The case for describing China as larger than Canada, India or the Soviet Union clearly rests in our greater interest in human beings than in square miles, tigers, polar bears and flies.
But China was smaller than the Soviet Union or even Russia by itself. Sen is simply repeating a lie.
...The non-predictive, non-prescriptive motivations are particularly important to bear in mind in understanding the role of “theories of value” in economics, particularly the labour theory of value.
The thing was nonsense. We understand that shitheads have a motivation to talk bollocks but we can exclude that bollocks from the realm of Accountancy & Economics & Actuarial or Management Science.
The labour theory can, in fact, be given various intrepretations related to the motivations involved.
So can the God theory or the Lesbian goat theory. Did you know that all these so called 'factories' and 'coal mines' and so forth are actually manned by Lesbian goats? Why are they being so pitilessly exploited? Is it because Manmohan Singh is secretly sodomizing the entire peasantry and proletariat of the world so incessantly that Capitalism has had to replace manpower with goatpower?
One is that of a predictive theory—in particular that of relative prices and distribution, and it encompasses the so-called “law of value’. Another is normative, providing a theory of entitlements. A third, which is frequently missed, is a way of discriminating between different features of production and exchange by focussing on elements of human involvement.'
Lesbian Goat theory has all these properties to the same extent as the Labour theory of Value.
Value, in Marx’s (1887) words, is ‘“‘a relation between persons expressed as a relation between things”.
only in the sense that a relation between lesbian goats and Neo-liberalism entails that shite.
The last interpretation is frequently taken to be “metaphysical”, as Joan Robinson (1964) puts it, describing it also as “‘a mere rigmarole of words” (p. 39). Why so? Because it
is nonsense. Value is a predicate. It isn't a relation. We say 'diamonds have value'. We don's say 'Necessity is the mother-in-law of Value. Charity used to date Value's ex but has now decided to marry a Lesbian goat.'
picks up from the description of production and exchange only some features as being specially significant, leaving the rest untouched? The discrimination makes the description arbitrary, as many critics have seen it. But, as I have already tried to argue, any description involves discrimination and selection, and the real question is the relevance of the selection process to the objectives of description.
To say 'to get to Victoria from Earls Court, take the District line' involves 'discrimination and selection'. But it is useful and 'Schelling focal'- i.e. is the 'natural' solution in the underlying coordination game. To say 'to get to Victoria from Earls Court, sodomize the ineluctability of the diagonalization of the Lesbian goat' also involves 'discrimination and selection'. But it isn't useful at all. Yet, that is what Sen is doing.
I have claimed elsewhere’* that the methodological issue can be more easily understood by looking at other descriptive statements in which the discrimination involved cannot be reduced entirely to predictive or prescriptive interests. Consider the statement “Michaelangelo produced the statue of David”. There is an obvious sense in which this would be accepted as
true based on Kripke rigid designators
a realistic description,
It is not a description.
despite its being informationally selective and the selection process not being primarily motivated by prediction or prescription. The production process in making the statue actually involved not merely Michaelangelo, but his helpers, a huge block of stone, chisels, scaffoldings, etc., but the description quoted focusses on Michaelangelo only as the most relevant bit of information. Note that the discrimination cannot be based on any marginal productivity consideration in the usual neoclassical sense. Without Michaelangelo, no statue, but without stone, no statue either! But this does not give the same status to the two statements “Michaelangelo produced the statue of David’ and “this stone produced the statue of David” (or “‘this stone quarry produced the statue of David’’).
Nothing wrong with saying that. But these are not descriptive statements. They are based on 'rigid designation'. To be descriptive they would be of the form 'Michaelangelo produced the sculpture of David which is x metres high and which weighs y kilograms and which looks like the picture below
The selection process reflects other motivations, in particular, that of capturing the source of the imagination displayed in the statue.
This has nothing to do with description
The labour theory of value, in its descrptive interpretation,
it has none. A theory is not a description.
shows a similar—but not the same—type of discrimination, focussing—in this case—
which case? Michaelangelo?
on the human effort directly and indirectly involved in the process of production and exchange of commodities.
Does Sen think the guy was working in a factory?
In assessing the labour theory of value as a description of production and exchange,
which is like assessing the cat as a description of the dog. A theory is not a description.
it is the cogency of this focus that has to be assessed,
there is zero cogency. The cat is not a dog. The theory is not a description.
including its ability, if any, to capture our interest compared with other methods of discrimination.
Why capture the interest of useless nutters?
Examining only the theory’s predictive success, or its normative relevance, will not even begin to provide
Sen with nutritious dog turds.
a complete assessment of the labour theory of value.
which is fully provided by saying the thing is dog shit.
And to call it metaphysical because it is informationally discriminating
in favour of theses which are 'beyond physics' and thus metaphysical, is perfectly justified.
Sen's analysis
involves a total failure to perceive the nature of description as an activity.
Briefly, the nature of this activity is to map what is observed to means of expression such that a description is available at some later date or to some third party. Some pathology, or previous instruction, may cause only some parameters to be observed and described. But this is not a matter of Choice. By contrast, an 'impressionistic' or otherwise aesthetic or polemical or other non-alethic account may involve choice.
I turn now to the other of the two great theories of value, viz., the utility theory.
Utility is a Tarskian primitive.
Again, utility descriptions have predictive and normative interest,
There are no utility descriptions. There are imputations of utility of a ipse dixit type.
but once again they do not exhaust the descriptive motivations.
They are wholly unrelated to them.
The joys and sufferings of human beings and their deprivations and fulfilments have interest of their own.
No. They are of no interest to anybody. Shitheads may pretend otherwise to virtue signal.
Even classical political economists were much concerned with related—though not identical—features in their interset in use value.'* It is this part of the descriptive motivation that has been least well served by recent developments of utility theory, most notably by the theory of revealed preference.
Which came out in 1938 when Sen was 5 years old. How is that 'recent'?
Focussing only on predicting behaviour, the richness of human psychology has been substantially ignored,
because Economists are ignorant of Psychology. You have to ignore what you haven't studied.
refusing to see anything in utility or happiness other than choice. As the “‘father’’ of revealed preference theory, Paul Samuelson, puts it: behaviour is to be “explained in terms of preference, which are in turn defined only by behaviour.”’
Which is all that businesses and governments are interested in. Crazy people, on the other hand, may want to discuss the Labour theory of value or the Lesbian Goat theory of value or the psychological richness involved in being incessantly sodomized by invisible Netan-Yahoos.
This has led to an approach that is—despite some predictive merits— remarkably mute about human joys and sufferings in which economics used to take a lot of interest.
Because it was useless and consisted only of scolding people rather than economising on the use of scarce resources.
The result is a descriptive impoverishment from many perspectives
Brown monkey whose people were starving just six years ago worries about 'descriptive impoverishment'! Yet, thanks to Western technology like TV cameras, descriptions now are better than ever!
, including—among others—normative relevance, since many of the common norms do relate precisely to the ignored descriptive features.'?
All the norms are dog shit which Sen must eagerly devour to overcome impoverishment.
... There has been the proposition that every factual statement involves implicit values.
not to mention invisible Lesbian goats.
In its original formulation and use, this claim, which I believe is false, may have played a challenging part in the dialectics with positivism. ‘“‘So you want facts only, no values, because they are illegitimate,” the argument ran, “but you can’t have facts without values. Touché!” But why must every factual statement involve values?
Invisible lesbian goats. They are very insistent on this point.
The basis of the claim seems to rest uneasily on the belief, which is correct, that any description involves some selection.
Which could be random and thus unconnected with human intentionality or bias or values or interests.
What is not correct is the further belief that the selection must be based on some explicit or implicit prescriptive criterion. The criteria used for selection, as I have been arguing, may be aimed at objectives other than prescription, e.g., catering to curiosity.
or sodomy or the ravenous devouring of dog turds. Since Economics is concerned with economising on the use of scarce resources, only descriptions which advance this objective are of interest to Economists.
... few concluding remarks. First, description inevitably involves selection.
and selection inevitably involves description. So what? If the thing is inevitable, how are economists concerned with it?
It can be usefully seen as a choice of a subset from a set of possible statements.
If there is no set, there can be no subset. 'Possible statements' does not have a well-defined extension. Thus it isn't a set.
It is not useful to see a thing as something which it can't be.
Second, truth is clearly not a sufficient condition for a description to be good.
So what? There is no necessary, sufficient, or other condition for anything to be anything save by arbitrary stipulation. But. by just such stipulation, Sen eats only dog turds.
Third, nor is it a necessary condition. Friedman’s defence of departures from truth in choosing factual assumptions is a special case of a general argument in favour of judging description by its usefulness.
No. There are no descriptions in Friedman. There is data- e.g. relating to the Monetary base or the Price level or the velocity of ciriculation. But data, by itself, is not description. What matters is whether Freidman's theory yielded the right predictions. In so far as the Phillips curve broke down- such was the case. But the prediction that inflationary pressure could be squeezed out purely by monetary policy without mass unemployment was soon belied. It turned out that Money wasn't truly exogenous.
Friedman considers only predictive usefulness, but there are other objectives which lead to the same conclusion. Samuelson’s assertion of the need to “‘jettison”’ models that contain “‘empirical falsities’” seems to take a very narrow view of description as an activity.
I suppose one could say that those with more knowledge of actual monetary policy were aware that Friedman't fairy tale was too good to be true. Still, the truth is, Friedman was actually the most successful 'normative' economist of his generation. He made it okay to laugh heartily at mass unemployment in the Rust Belt while pulling in big bucks from a job in Finance.
Fourth, Friedman’s attack on traditional notions of realism is based on a misunderstanding. His own suggestion about judging realism by ‘“‘the test by prediction”? seems to confound usefulness of a description with its realism.
The thing didn't matter greatly. Friedman & Hayek prevailed because those who embraced their ideas and values did better for themselves thereby. Bleeding Heart Keynesianism put money in nobody's pocket ; it was simply a pain the arse.
A description of something can be a good one to give without being a good description of that thing.
What Sen means is 'sometimes, it is better for you to tell a lie rather than the truth'.
Fifth, description can be motivated by predictive interest or by prescriptive interest, but it may also have other motivations, and to confine attention only to predictive or prescriptive interest impoverishes the traditions of descriptive economics.
Lying your head off may make you rich. At the very least, as a liar, you can claim to have become wealthy.
Finally, such impoverishment can be seen in recent developments in many different fields of economics, varying from interpretations of theories of “value” to the measurement of inequality and poverty.
People say 'poor Professor X. As an undergrad, he appeared so promising. Now everybody thinks he is a stupid liar. His books and papers are considered very poor quality work'. This type of 'impoverishment' can be defeated, Sen believes, by telling even stupider, even more monumental, lies.
The richness of descriptive motivations seems to have been sacrificed for consistency within an arbitrarily narrow conception of these motivations, or—to be more exact—for consistency respectively within two arbitrarily narrow conceptions of motivations.
The richness of fantasist and the luxuriance of the liar has been sacrificed by professional people who get rich by discoveritng and reporting the truth.
This, I have argued, has been an unequal exchange: confounding the nature of description as an activity and unnecessarily weakening the theoretical underpinning of many legitimate and useful activities in the social sciences.
Normative economics is based on telling stupid lies. Thus Sen, attacking Thatcher a couple of years after this paper was published, said that the UK could suffer a big famine under her leadership. His colleagues discouraged this claim. Brown monkeys should stick to denouncing other brown monkeys. Thus they welcomed his outrageous libel that Manmohan Singh was trying to force horribly diseased, wholly illiterate, Indian people into modern factories and offices so as to turn Indian into an economic super-power. It was actually the 'Dobb-Sen' thesis which had envisaged confiscating productivity gains from the proletariat so as to invest in the capital good sector.
A fair description of Sen's entire oeuvre would leave us with no choice but to declare him a useless, stupid, ignorant liar and virtue signalling hypocrite. We are welcome to enrich this description, in the manner he himself endorses, by making the ipse dixit stipulatiton that he eagerly devours dog turds.
