Monday, March 30, 2015

The Two Epistemological Ways to Interpret the Labour Theory of Value

The labour theory of value can be expressed as the following proposition:
“The labour theory of value (LTV) is an economic theory of value that states that the economic value of a good or service is determined by the total amount of socially necessary labour required to produce it.”
There are two ways in which, epistemologically speaking, we could interpret this proposition, as follows:
(1) as a synthetic a posteriori proposition (= an empirical proposition), or

(2) as an analytic a priori proposition.
If we interpret it as (1) an empirical proposition, then the LTV is not necessarily true, but contingent and known as true only a posteriori by experience, empirical evidence, and inductive argument. As a matter of historical fact, Marx took the LTV from Ricardo, who in turn took it from Adam Smith, and these men seem to have taken it as an empirical statement.

But now any Marxist who defends the LTV as an empirical proposition faces the demand to prove by experience, empirical evidence, and inductive argument that such a type of value really exists. They have to show us how abstract socially necessary labour time (SNLT) can be actually and meaningfully defined as a real measure of heterogeneous labour. They have to show how to calculate the abstract socially necessary labour time (SNLT) unit values of commodities. They need to show how these SNLT values map onto, or correspond to, the “natural” or “true” exchange values or prices of commodities, and identify real world prices that are direct examples of such SNLT values.

Alternatively, if we interpret the LTV as (2) an analytic a priori proposition, then the LTV is necessarily true and known a priori (not by any empirical evidence), but strictly speaking as a proposition it is a mere definition or tautologous statement. It is an analytic statement like:
(1) all bachelors are unmarried.
Even if there were actually no bachelors in the world, this proposition would still be necessarily true, but it is a mere analytic truth, that is, a mere definition or tautologous statement.

Experience teaches us of course that there are things in the world that are men who are not married, and so can be classified as unmarried men or bachelors.

But we make a deep mistake if we do not scrutinise analytic statements carefully to see if they are coherent, consistent, meaningful and useful, and – above all – whether they actually refer to things that have real existence and empirical relevance. The analytic proposition “all bachelors are unmarried” fulfils these criteria. It is a coherent and useful analytic proposition that can be used to classify objects in the real world into a class.

But we could for example propose the following analytic statements:
(1) all unicorns have horns.

(2) All dragons are able to fly and breathe fire.
It is obvious that the real world does not contain unicorns or dragons (at least we have no rational reason to think so). But these are still valid and true analytic propositions because there are worlds of human fiction where these things are imagined to exist, e.g., the dragons in the Lord of the Rings novels and films. Now an imaginary world is analogous to a highly artificial model that does not explain or relate to anything in the real world.

The propositions above about dragons and unicorns are only useful and relevant as analytic statements referring to a purely imaginary or hypothetical world in our minds – say, the world of human fiction or fantasy writing.

If the Marxists wish to interpret the LTV as analytic a priori proposition, then they could defend it as a merely empirically-empty and tautologous definitional statement. But they would still face the tremendous hurdle of proving that their analytic definitions are something more than analytic statements referring to a purely imaginary or hypothetical world in our minds, on a par with “All dragons are able to fly and breathe fire.”

That is to say, they still face the same severe empirical challenges that they would face if they choose to interpret the LTV empirically.

If Marxists cannot do this, and they retreat to a defence of the LTV as an analytic truth, it seems to me the only sensible way to treat the LTV is as a purely imaginary or hypothetical, useless proposition that adds nothing to economic knowledge. It has no explanatory power in the analysis of a real world economy. As a postulated phenomenon, it cannot be found in the real world and has no discernible causal power as a factor in a real world economy, unlike real higher-level emergent properties like the general state of expectations or aggregate demand.

And, finally, what would even be the point of asserting the LTV as some analytic concept? It would be nothing but a pointless and worthless idea, with no explanatory or causal power, tacked on to real economic science, which is at least a body of empirical truths and defensible, empirically-tested theories.

The LTV is, quite simply, on a par with the Austrian and neoclassical concept of the “natural rate of interest” or Milton Friedman’s “natural rate of unemployment”: we are dealing with concepts only conceivable in wholly higher-level abstract models that are so unrealistic and so remote from, and so irrelevant to, the real world that they are nothing but fictions.


  1. Not all synthetic a posteriori statements are practically testable and "provable". In fact, scientific statements, especially if they involve theoretical terms which refer to unobservable entities, are never tested in isolation, but face the tribunal of experience as an ensemble of statements. This has been common knowledge in the philosophy of science since the times of Duhem, or, at latest, of Quine. What you are proposing is 1920s Vienna style logical positivism, which even its founders were forced to abandon.

    The tasks for Marxists that you proposed cannot be solved. But Marx's theory of value has some implications which can be tested; if the the results are positive, this (potentially) counts as positive evidence for the theory, if not, the it counts as negative evidence.


    1. This is not logical positivism at all. I am not asserting that if something cannot be empirically verified, it is literally nonsense.

      Now there are some synthetic a posteriori statements that cannot be tested today because we lack the technology, e.g., "there is carbon-based life on the 2nd planet from the sun of the Zeta 2 Reticuli system." But we clearly know in principle how to verify and falsify them.

      Contrast that with the LTV. We are not dealing far distant planets or dark matter or electrons -- but observable human labour and observable prices.

      Yet you can't even adequately explain exactly how to define abstract socially necessary labour time and show me how such a value determines a natural market price.

      In fact, I've challenged you to tell me what it would take to falsify the LTV for you and you can't even tell me.

    2. Also, you haven't even answered the question:

      Is the LTV merely (1) an analytic proposition or (2) an empirical statement?

    3. LK, it's neither 1 nor 2. It's something that is testable only with additional and only partially true assumption that prices right now are equal to values. Same problem happens to all static price theories in a dynamic world, by the way.

    4. For christ's sake -- it's neither empirical nor even analytic?! Then pray tell what mysterious epistemological category it's in? Unfalsifiable metaphysics?

      What you are saying is that the LTV is just like the natural rate of interest: a totally fictitious concept in an imaginary world in people's heads, with zero relevance to the real world.

    5. It's a building block of an empirical theory. I don't remember how it's classified by philosophy. You're the expert philosopher, not me! :)

    6. The LTV is obviously a set of empirical propositions. The core of the theory cannot be tested on its own. The implications can. A fairly standard Lakatosian situation.


    7. What is the core of the theory in your view?

      That all heterogeneous concrete labour can be reduced to a meaningful well-defined abstract SNLT and that this is in turn equivalent to the "natural" price/exchange value of commodities?

      If you are saying that these things "cannot be tested" that is a blatant admission of how anti-empirical and theological your theory is. And I do not subscribe to Lakatosian theory of science, so this doesn't cut any ice with me.

    8. No, there's no equivalence like that. We've already established this some time ago. The core are the definitions of value, abstract labor, socially necessary labor time, value of labor power, production price etc. The testable consequences I can think of concern long-term capitalist development. I gave a list of what I think could be used to test the theory in a previous thread.

      "Value" is simply a theoretical term which denotes an unobservable. But you can observe its effects. The trick is to derive predictions about effects that are not derivable from any of the rival theories.

      I don't really care to what philosophy of science you subscribe, the thing is that exactly the same problems occur in other disciplines. Try testing any physical law as an isolated statement.


  2. LK, we're two different anonymous, my excuses for the confusion... and, yes, ignoring heterogeneous labor is another approximation, and i agree with you that Marx seems to have forgotten that. This general tendency to confuse approximations with exact statements is a step backward from Ricardo. And Ricardo is a step backward from empirically support theories like mark up pricing that you've nicely documented in your blog.

  3. There is actually a logical solution to the LTV, which is based on Sraffa's standard commodity, and implies that long term prices are determined by labor commanded as in Adam Smith. So no, it is incorrect to compare the LTV with the logically flawed notion of the natural rate of interest. My post on that

    1. That looks like a pure trick. Couldn't one argue prices are proportional to "lumber commanded"?

  4. Yes, we know you know the difference between an analytic a priori and a synthetic a posteriori. Marx's value theory has concrete definitions that it structures into a synthetic system that generates empirical statements about the world. In other words, like literally every science, it incorporates both types of proposition. I've even provided you with a very important recent piece of empirical scholarship on the matter.

    But no, you've got your head down and you're cranking out post after post just as quickly as you can to show how Right And Also The Winner you are, grasp of the subject matter be damned.

    It's frankly hard to watch, because I really do think you're a lot smarter than this.