Thursday, March 31, 2016

Alan Turing’s “Computing Machinery and Intelligence”

Ken B, take note!

Since I am getting flack for being skeptical about the Turing test, let me review Alan M. Turing’s original paper “Computing Machinery and Intelligence” (1950).

As a matter of pure historical interest, one of the first people to imagine intelligent machines was the 19th century novelist Samuel Butler in the novel Erewhon (London, 1865), which is actually cited by Turing in his bibliography of this paper (Turing 1950: 460). A case of life imitating art?

Anyway, I divide my post below into two parts:
I. Turing’s Paper “Computing Machinery and Intelligence”: A Critical Summary

II. Critique of the Turing Test.
Turing’s paper is based on the type of truly embarrassing and crude behaviourism that was fashionable in the 1940s, and, oh my lord, it shows.

So what was behaviourism? I quote from the Internet Encyclopedia of Philosophy:
“Behaviorism was a movement in psychology and philosophy that emphasized the outward behavioral aspects of thought and dismissed the inward experiential, and sometimes the inner procedural, aspects as well; a movement harking back to the methodological proposals of John B. Watson, who coined the name. Watson’s 1913 manifesto proposed abandoning Introspectionist attempts to make consciousness a subject of experimental investigation to focus instead on behavioral manifestations of intelligence. B. F. Skinner later hardened behaviorist strictures to exclude inner physiological processes along with inward experiences as items of legitimate psychological concern.”
Hauser, Larry. “Behaviorism,” Internet Encyclopedia of Philosophy
Ouch! The very essence of behaviourism was to abandon the study of internal mental or biological states of human beings and human consciousness to focus on outward “behavioural manifestations of intelligence.” Behaviourism had no interest in the internal explanation of human mental states and intelligence, and, instead, focused on external signs of it.

Behaviourism led to some real intellectual disasters in 20th century social sciences and philosophy. It shows up all over the place.

For example, B. F. Skinner’s Verbal Behavior applied the behaviourist paradigm to linguistics in a deeply flawed manner, which was brought out by Noam Chomsky’s now famous 1959 review of that book (Schwartz 2012: 181).

Even the analytic philosopher Willard Van Orman Quine’s misguided attempt to deny a valid distinction between analytic and synthetic propositions is a legacy of crude verbal behaviourism.

I. Turing’s Paper “Computing Machinery and Intelligence”: A Critical Summary
Turing divided his paper into the following sections:
(1) The Imitation Game
(2) Critique of the New Problem
(3) The Machines concerned in the Game
(4) Digital Computers
(5) Universality of Digital Computers
(6) Contrary Views on the Main Question
(7) Learning Machines.
Let us review them one by one.

(1) The Imitation Game
Turing proposes to answer the question: “Can machines think?” He rightly notes that any sensible discussion should begin with a definition of “machine” and “think” (Turing 1950: 433).

Unfortunately, no such definition is given. Instead, Turing rightly notes that a serious definition would not simply rely on an opinion poll of what people think, but then the issue is thrown aside. In place of a definition, Turing proposes the “Imitation Game.”

In this game, we have three players, A, B, C.

A is a machine, B a human being, and C a person who asks questions indirectly of A and B (who remain hidden from C). If C asks questions of A and B at length and cannot determine who the computer is and who the human is, then the machine is deemed to have passed the test (Turing 1950: 433–434).

At this point, it should be obvious that the “Imitation Game” is the Turing Test.

(2) Critique of the New Problem
Since the machine or computer remains hidden, Turing emphasises that this is a test of behaviour or “intellectual capacities,” not, say, external appearance (Turing 1950: 434).

The “best strategy” for the machine is “to try to provide answers that would naturally be given by a man” (Turing 1950: 435).

At this point, Turing raises the crucial question:
“May not machines carry out something which ought to be described as thinking but which is very different from what a man does? This objection is a very strong one, but at least we can say that if, nevertheless, a machine can be constructed to play the imitation game satisfactorily, we need not be troubled by this objection.” (Turing 1950: 435).
Just look at how Turing flips this question off.

Since Turing has no interest in defining “intelligence” and commits himself to a crude behaviourist test, his nonchalant attitude here is understandable.

But it remains a devastating problem with his whole approach and, moreover, undermines the worth of the Turing test.

For the question Turing raises – can a machine “carry out something which ought to be described as thinking but which is very different from what a man does?” – cannot be evaded. It is at the heart of the problem.

Let us propose two preliminary definitions of “intelligence” as follows:
(1) information processing that takes input and creates output that allows external behaviour of the type that animals and human beings engage in, and

(2) information processing that takes input and creates output that is accompanied by the same type of consciousness that human beings have.
Now a behaviourist would be interested in (1), but can ignore (2).

But (2) is the philosophically and scientifically interesting question.

(3) The Machines concerned in the Game
Turing now clarifies that by “machines” he means an “electronic computer” or “digital computer”: he only permits digital computers to be the machine in his Turing Test (Turing 1950: 436).

It is also explicit in the paper as well that Turing envisaged future computers as being sophisticated enough to pass the test, not the computers of his own time (Turing 1950: 436).

(4) Digital Computers
In this section, Turing explains the nature and design of digital computers.

But to make a computer mimic any particular action of a human being, the instructions for that action have to be carefully programmed (Turing 1950: 438).

(5) Universality of Digital Computers
Turing discusses discrete state machines (Turing 1950: 439–440), and points out that digital computers can be universal machines in the sense that one computer can be specifically programmed to compute a vast array of different functions (Turing 1950: 441).

Turing now returns to his main question:
“It was suggested tentatively that the question, ‘Can machines think?’ should be replaced by ‘Are there imaginable digital computers which would do well in the imitation game?’ If we wish we can make this superficially more general and ask ‘Are there discrete state machines which would do well?’ But in view of the universality property we see that either of these questions is equivalent to this, ‘Let us fix our attention on one particular digital computer C. Is it true that by modifying this computer to have an adequate storage, suitably increasing its speed of action, and providing it with an appropriate programme, C can be made to play satisfactorily the part of A in the imitation game, the part of B being taken by a man?’” (Turing 1950: 442).
(6) Contrary Views on the Main Question
Unfortunately, Turing’s answer to the question whether computers can think is just as nonchalant as in section 1:
“It will simplify matters for the reader if I explain first my own beliefs in the matter. Consider first the more accurate form of the question. I believe that in about fifty years’ time it will be possible to programme computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 per cent. chance of making the right identification after five minutes of questioning. The original question, ‘Can machines think?’ I believe to be too meaningless to deserve discussion. Nevertheless I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.” (Turing 1950: 442).
The very question “Can machines think?” is dismissed by Turing as “too meaningless to deserve discussion.”

There is not even any attempt to properly delineate or define the sense in which the word “intelligence” might be understood.

For Turing, all that matters is that the computer can successfully play the imitation game.

Turing then turns to objections which might be proposed to the idea that computers can think:
(1) The Theological Objection
This is the objection that human beings have an immaterial essence or soul that makes them intelligent. Turing dismisses this.

(2) The ‘Heads in the Sand’ Objection
This is really nothing more than the objection that machines being able to think is a horrible idea. Turing again responds that such an emotional response will not do.

(3) The Mathematical Objection
Turing considers the limitations of discrete-state machines in relation to Gödel’s theorem.

Turing replies to this by pointing out that the Imitation Game’s purpose is to make a computer seem like a human, so that incorrect answers or the inability to answer logical puzzles wouldn’t necessarily be a problem.

(4) The Argument from Consciousness
It is here we come to what should be the most interesting part of the paper.

Turing quotes a critic who rejects the idea that machines can ever equal the human brain:
“This argument [sc. from consciousness] is very well expressed in Professor Jefferson’s Lister Oration for 1949, from which I quote. ‘Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain—that is, not only write it but know that it had written it. No mechanism could feel (and not merely artificially signal, an easy contrivance) pleasure at its successes, grief when its valves fuse, be warmed by flattery, be made miserable by its mistakes, be charmed by sex, be angry or depressed when it cannot get what it wants.’” (Turing 1950: 445–446).
Now this goes too far in demanding that the machine needs to directly experience human emotion, because it is, I imagine, possible for a human being to not feel emotion from brain disorder but still be conscious and intelligent.

Nevertheless, the demand that a machine would have to be fully conscious like us to be the equal of the conscious intelligent human mind is sound.

What is Turing’s response? Turing says that Jefferson demands that “the only way by which one could be sure that a machine thinks is to be the machine and to feel oneself thinking” – that is, we need solipsistic proof (Turing 1950: 446).

But that is a straw man and misrepresentation of what Jefferson said. Jefferson said not that he would need to be the machine to be convinced that it is the equal of the human brain, but that it must have the same conscious life as the human brain. Turing’s dismissal of this will not stand.

Turing then argues that if a computer could actually write sonnets and answer aesthetic questions about it, then it could pass the Turing Test and meet Jefferson’s demand (Turing 1950: 447).

(5) Arguments from Various Disabilities
Here Turing replies to critics who argue that, no matter how sophisticated a computer is, there must always be something new that they cannot do that humans can do.

Turing responds to this by saying that computers with more memory and better and better programs will overcome such an objection (Turing 1950: 449).

(6) Lady Lovelace’s Objection
This stems from objections made by Lady Lovelace to Babbage’s Analytical Engine.

Lovelace essentially said that such machines are bound by their programs and cannot display independent or original behaviour or operations (Turing 1950: 450).

Turing counters by arguing that sufficiently advanced computers might be programmed to do just that by learning (Turing 1950: 450).

(7) Argument from Continuity in the Nervous System
Here we get an interesting objection: that human brains have nervous systems and so what they do arises from a different biological substrate from that of electronic computers:
“The nervous system is certainly not a discrete-state machine. A small error in the information about the size of a nervous impulse impinging on a neuron, may make a large difference to the size of the outgoing impulse. It may be argued that, this being so, one cannot expect to be able to mimic the behaviour of the nervous system with a discrete-state system.

It is true that a discrete-state machine must be different from a continuous machine. But if we adhere to the conditions of the imitation game, the interrogator will not be able to take any advantage of this difference.” (Turing 1950: 451).
So Turing here dismisses the objection by saying that this doesn’t matter provided that the computer can fool an interrogator.

But this behaviourist obsession only with external output will not do. If we wish to answer the question “can computers attain the same conscious intelligence as people,” there must be a serious attempt to answer it. There is none.

(8) The Argument from Informality of Behaviour
The argument here is that human behaviour encompasses a vast range of activities and choices that cannot be adequately listed or described in rule books or programs.

Turing replies that nevertheless overarching general rules of behaviour are sufficient to make machines appear as humans (Turing 1950: 452).

(9) The Argument from Extra-Sensory Perception
In a quite bizarre section, Turing raises the possibility that human beings have extra-sensory perception, such as telepathy, clairvoyance, precognition and psycho-kinesis (Turing 1950: 453), that machines can never have.

Turing even states that “the statistical evidence, at least for telepathy, is overwhelming” (!!) (Turing 1950: 453). This would pose problems for the Turing Test apparently, though Turing’s argument is rather confused.

One solution, Turing suggests, is that we would need to have a “telepathy-proof room” in Turing Tests! (Turing 1950: 453).
(7) Learning Machines
In the final section, Turing suggests that the human mind is fundamentally a mechanical system, though not a discrete-state machine (Turing 1950: 455).

Turing thought that by the year 2000 there would be a definitive answer to question whether machines can regularly pass the Turing test, given a tremendous increase in memory and complexity of programming (Turing 1950: 455). Turing suggests that the first step is to create a computer that can simulate the answers of children and be a type of learning machine (Turing 1950: 456–459).

II. Critique of the Turing Test
Let us return to the two definitions of “intelligence” as follows:
(1) a quality in human digital computers in which information processing takes input and creates output that allows action or sentences, or external behaviour of the type that animals and human beings engage in, and

(2) information processing that takes input and creates output that is accompanied by the same type of consciousness that human beings have.
If we define “intelligence” in sense 1, then of course you can say that computers are “intelligent,” and not just computers that pass the Turing Test.

But if we focus on sense (2) – which is the philosophically and scientifically interesting question – it is not at all clear that computers have or can ever have the same *conscious* intelligence that human beings have.

Now the behaviourist Turing Test obviously influenced the functionalist theory of the mind, which holds that functional mental states may be realized in different physical substrates, e.g., in a computer. This freed functionalist psychologists and artificial intelligence researchers from a strict dependence on neuroscience. Since, for the functionalists, mental states are abstract processes that can be created in multiple physical systems, AI researchers using functionalism were free to study mental processes in a way that did not reduce their disciples to mere study of brain neuroscience and its physics and chemistry.

Those AI researchers following Turing adopted a top-down “Good Old Fashioned AI” (GOFAI) (or symbolic AI) research program and really thought that they could create artificial intelligence as rich as human intelligence. But their attempts to create a human-level artificial intelligence ended in miserable failure, and manifestly did not succeed. Of course, we did get some very useful technology out of their research, but none of these computers can remotely approach the full level of human intelligence.

Where did AI go wrong?

Let me start to answer that question by reviewing John Searle’s Chinese Room argument.

This was first presented in 1980, and Searle imagines a room in which he is present, with an opening through which paper can be pushed inside or outside. Searle receives papers through the slot in the room. On the paper are symbols which Searle can look up in a complex system that allows him to match the symbols and then to write another set of symbols on a paper which he can then pass out of the room. The system is a complex set of algorithms that provide him with instructions for producing new symbols. The symbols that Searle manipulates are in fact written Chinese and the responses will be intelligent answers to Chinese questions. Searle argues that, because he does not understand Chinese, a mere process that manipulates symbols with a formal syntax can never actually understand the meaning of the symbols. An understanding of real meaning, in other words, is impossible for such a system. We need a human mind with consciousness and intentionality for that.

This argument deployed by Searle is directed against Strong AI’s classical computational theory of mind, not just against Turing’s “Imitation Game,” for as we have seen Turing did not even care or address the question whether machines were conscious like human beings. According to the classical computational theory of mind, a sufficiently complicated program run on a digital computer could do what the human mind does, and this program itself would also be a mind, with the same cognitive states (consciousness, sensation etc.) that a human mind has.

Searle believes that his Chinese room argument shows that Strong AI is completely mistaken, and that mere algorithmic manipulation of symbols with syntax can never produce a conscious mind with perception, sensation and intentionality. Searle believes that he has shown this because the symbols manipulated in the Chinese room could be any type of information (e.g., text of any language or auditory or visual information), yet the person or system that manipulates them does not understand their meaning. In Searle’s view, the computation is purely syntactic—there is no semantics (Boden 1990: 89). So Searle argues that, if the person manipulating the symbols does not understand their meaning, then no computer can either if it only uses a rule-governed symbol manipulation (Searle 1980: 82).

Searle also criticises the Turing Test in the Chinese Room argument, since the room, if one imagined it as a computer, could pass the Turing test but still have no understanding. The mere appearance of understanding, then, is no proof of its existence. This, Searle argues, is a serious flaw in Strong AI, because the Turing Test is deeply dependent on a mistaken behaviourist theory of the mind (Searle 1980: 85). Turing might reply that he does not even care if the computer has understanding or consciousness, and so is not even concerned with the question whether a computer can attain “intelligence” in sense (2) above.

Searle has also criticised the connectionist theory in a modified version of the Chinese Room argument, which he calls the Chinese Gym (Searle 1990: 20–25).

The responses to Searle are various, but connectionist critics of Searle argue that we cannot predict what emergent properties might arise in computers with vast amounts of parallel processing and vector transformations (Churchland 1990: 30–31).

But, of course, unless science properly understands precisely how consciousness emerges from the brain, there is no definitive answer.

For John Searle, the biological naturalist theory of the mind is the best one we have, and human minds are a type of emergent physical property from brains and necessarily causally dependent on the particular organic, biological processes in the brain. Synthetic digital computers cannot attain consciousness because they lack the physically necessary biological processes.

Another point is that, if we say that the human brain involves information processing, then surely it must be the case that a very different type of information processing goes on in the brain compared with that which occurs in digital computers, and clearly many aspects of the human mind are not computational anyway.

Finally, if a computer passes a Turing Test, then all it demonstrates is that it is possible to simulate the verbal behaviour of human beings. It does not follow a computer can have conscious intelligence in sense (2) above.

My final point is mainly flippant.

Recently, another amusing chapter in the story of AI happened: Microsoft’s AI Chatbot called “Tay.”

Microsoft launched its chatbot Tay on Twitter, an AI program that writes tweets by learning from people who chat to it, and so supposedly learning to write Tweets and simulate conversations that sound more and more like those of a real human being.

It seems that certain people trolling this bot began frequently talking to it about highly – shall we say? – controversial things, and influencing the type of Tweets it was writing.

Tay was supposed to act like a teenage girl on Twitter. How did that experiment in AI go?

Within a day, Tay started spewing forth Tweets:
(1) that denied the truth of the Holocaust;

(2) that expressed support for Nazism and asserted that Hitler did nothing wrong, and

(3) that called for a genocide.
Not exactly another success for AI!

This is all described by the YouTube personality Sargon of Akkad in the video below.

N.B. Some bad language in the video and things in bad taste!

Boden, M. A., 1990, “Escaping from the Chinese Room,” in Boden, M. (ed.), The Philosophy of Artificial Intelligence. Oxford University Press, Oxford. 89–104.

Chomsky, Noam. 2009. “Turing on the ‘Imitation Game,’” in Robert Epstein, Gray Roberts and Grace Beber (eds.), Parsing the Turing Test: Philosophical and Methodological Issues in the Quest for the Thinking Computer. Springer, Dordrecht. 103–106.

Churchland, Paul, Churchland, Patricia, 1990, “Could a Machine Think,” Scientific American 262.1 (January): 26–31.

Churchland, Paul M. 2009. “On the Nature of Intelligence: Turing, Church, von Neumann, and the Brain,” in Robert Epstein, Gray Roberts and Grace Beber (eds.), Parsing the Turing Test: Philosophical and Methodological Issues in the Quest for the Thinking Computer. Springer, Dordrecht. 107–117.

Hauser, Larry. “Behaviorism,” Internet Encyclopedia of Philosophy

Schwartz, Stephen P. 2012. A Brief History of Analytic Philosophy: From Russell to Rawls. Wiley-Blackwell, Chichester, UK.

Searle, J., 1980, “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3: 417–458.

Searle, J., 1990, “Is the Brain’s Mind a Computer Program,” Scientific American 262.1 (January): 20–25.

Turing, Alan. 1936. “Computable Numbers, with an Application to the ‘Entscheidungsproblem,’” Proceedings of the London Mathematical Society 42.2: 230–267, with 43 (1937): 544–546.

Turing, Alan M. 1950. “Computing Machinery and Intelligence,” Mind 59.236: 433–460.

Wednesday, March 30, 2016

George Selgin versus David Graeber on the Origin of Money

The blogosphere has recently seen a new debate on the origin of money between George Selgin and David Graeber:
Ilana E. Strauss, “The Myth of the Barter Economy,” The Atlantic, February 26, 2016

Though not by Selgin or Graeber, I gather that this article set off the debate.

George Selgin, “The Myth of the Myth of Barter,” Alt-M Ideas for an Alternative Monetary Future, March 15, 2016.

This is Selgin’s original critique of Graeber.

George Selgin, “Graeber, Once More,” Alt-M Ideas for an Alternative Monetary Future, March 24, 2016

After some debate on Twitter, Selgin posted this reply.
For those of you who don’t know, George Selgin is a monetary economist (whose research includes banking theory, monetary history, and free banking, among other things), and David Graeber is an anthropologist and author of the fascinating book Debt: The First 5,000 Years (2011).

To some extent, we are simply revisiting the debate that David Graeber had with Robert Murphy had in 2012.

I have intellectual respect for both George Selgin and David Graeber. For example, Selgin’s paper “Those Dishonest Goldsmiths” is an excellent refutation of the Rothbardian cult on their history of the origin of fractional reserve banking in England.

But, on this issue, I get the impression that people here are talking past one another. I offer this in the spirit of constructive and friendly criticism of both George Selgin and David Graeber, and hope it is taken in that way.

Let me begin with a positive assessment of Selgin’s analysis.

Selgin argues the following:
(1) that just because there is no modern anthropological evidence for pure barter economies, it does not necessarily follow that these things did not exist in the distant past before money was invented.

(2) Selgin complains:
“What I do deny, and vigorously, is anthropologist David Graeber’s claim that the existence of gift economies undermines ... ‘the entire discourse of economics.’”
(3) there is an irrational hatred of capitalism from some left-wing people, who want to blame slavery and imperialism on the use of money, and even paint “Adam Smith … as an enabler of slavery and imperialism.”

(4) that human societies can have quantified debts in terms of commodities owed without money.

(5) that, if we define value as subjective value, then money is not an objective measure of value, as the Austrians argue. It is also true that, generally speaking, during exchanges one person subjectively values the good he receives more highly than the good he parts with. Exchange is not an exchange of equivalents in that sense when we consider subjective value.
I can grant George Selgin all these points. In fact, I do. He is essentially correct on these points.

Selgin is a highly intelligent man, and he has scored hits against left-wing people here.

On (2), of course the refutation of simplistic models of money emerging from pure barter in neoclassical textbooks doesn’t refute all of neoclassical economics. Such a claim is absurd.

On (3), yes, it is the worst sort of left-wing irrationality to blame slavery and imperialism on money or capitalism.

On (5), I also appreciate Selgin’s complaints about the Aristotelian view of exchange:
“The idea that money is a ‘measure of value,’ like the related idea that exchanges are necessarily exchanges of equivalents, is among the hoariest of economic fallacies. It plays a prominent part in Aristotle’s economics — and, not coincidentally, in Aristotle’s condemnation of all sorts of ‘capitalist’ activity. Smith himself, in subscribing to a modified labor theory of value, was unable to break free of it. It is more than a little ironic that Graeber, in flinging all sorts of undeserved criticism at Smith, cleaves to him when it comes to his one indisputable mistake.

The notion that money is a ‘measure of value’ is but a particular instance — albeit one that has managed to linger on in some economics textbooks — of the mistaken belief that economic exchanges are exchanges of equivalents.”
George Selgin, “The Myth of the Myth of Barter,” Alt-M Ideas for an Alternative Monetary Future, March 15, 2016.
One could also add: Marx took over that same idea of exchange as necessarily exchanges of equivalents from the Classical economists and Aristotle, and it is a major reason for the train wreck of illogic and flawed argument in Chapter 1 of volume 1 of Capital.

We get to the crux of Selgin’s counterargument here:
“My concern, though, isn’t with Graeber’s sweeping condemnation of modern economics, or of the economic arrangements for which modern economists are supposedly to blame. It’s with his particular claim that there’s no merit in Smith’s account of the origin of money, or in the later accounts of other economists, including Carl Menger. Despite what these economists have argued, money couldn’t have grown out of barter, Graeber insists, because the ‘fabled land of barter’ that these accounts posit never existed.”
George Selgin, “The Myth of the Myth of Barter,” Alt-M Ideas for an Alternative Monetary Future, March 15, 2016.
Unfortunately, it seems to me that this misunderstands Graeber’s theory.

Graeber, as far as I am aware (I could be wrong), does not say that there is absolutely no merit of any kind to the views of Adam Smith or Carl Menger.

Graeber does not deny that money in some historical circumstances can emerge from barter, especially in long distance trade.

Right on p. 75 of Debt: The First 5,000 Years (2011), Graeber says:
“Throughout most of history, even where we do find elaborate markets, we also find a complex jumble of different sorts of currency. Some of these may have originally emerged from barter between foreigners: the cacao money of Mesoamerica and the salt money of Ethiopia are frequently cited examples. Other arose from credit systems, or from arguments over what sort of goods should be acceptable to pay taxes or other debts. Such questions were often matters of endless contestation.” (Graeber 2011: 75)
This statement requires that, yes, there is some merit to the theories of Smith and Menger, but suitably qualified.

The point is, however, that the Smith, Menger and neoclassical “barter origin” theory of money is not universally true and cannot be in its classical form a correct theory of the origin of money. It’s flawed.

What is needed is an eclectic theory that takes careful account of history and anthropology.

Moreover, even Selgin himself concludes that “notwithstanding the fact that credit is older than barter, Smith’s theory is, after all, not all that far removed from the truth.”

So if Selgin has asserted that
(1) the empirical evidence strongly suggests that gift exchange and debt–credit exchanges without money long preceded the invention of money in human history, and

(2) the classical and orthodox barter spot trade theory of the origin of money needs revision.
then it follows directly that Selgin has conceded something significant to his opponents.

Selgin states:
“So, how true is Graeber’s account, and just how fatal is it to the "fable" that economists like to tell? For answers, we need look no further than the evidence Graeber himself supplies. For on close inspection, that evidence itself suffices to show that, notwithstanding the fact that credit is older than barter, Smith's theory is, after all, not all that far removed from the truth.

A paradox? Nothing of the sort. The simple explanation is that, while subtle forms of credit or outright gift giving may suffice for affecting exchanges within tightly-knit communities, exchange within such communities hardly begins to take advantage of opportunities for specialization and division of labor that arise once one allows for trade, not just within such communities, but between them, that is, for trade between or among strangers. One need only recognize this simple truth to resuscitate Smith's theory from Graeber’s seemingly fatal blow. Simple forms of credit may come first; but such credit only goes so far, because it depends on a repeated interaction, and the trust that such interaction both allows and sustains.”
George Selgin, “The Myth of the Myth of Barter,” Alt-M Ideas for an Alternative Monetary Future, March 15, 2016.
At this point, Selgin cites with approval Graeber’s own views that barter trade, historically speaking, was probably far more important between ethnic groups and different communities, particularly in long distance trade.

Selgin continues:
“The question is, what did Smith really ‘imagine’? His story of the butcher and the baker notwithstanding, his reference to pastoral societies makes it perfectly evident that he understood the difference between conduct among ‘villagers’ and conduct among strangers. His theory of the origins of money ought to be understood accordingly. It is a theory of how, when opportunities for trade arise among strangers, bringing with them further scope for the division of labor, trade will be ‘choked and embarrassed’ if it must occur by means of barter, but will cease to be so once barter gives way to the employment of money. In portraying such cases as exceptions to the rule that ‘credit’ proceeds barter, Graeber simply fails to understand that such ‘exceptions’ are all that matters in assessing Smith’s theory.” ….

In short, a generous reading of Smith, far from making him out to be a right bungler when it comes to matters ethnographic, yields a relatively sophisticated view, according to which kinship and ‘credit’ first predominate, but then give way, as strangers meet, first to barter, but eventually to monetary exchange, which in turn allows for the growth of commerce, which ends up reducing the role of kinship and kin-based credit relationships.”
George Selgin, “The Myth of the Myth of Barter,” Alt-M Ideas for an Alternative Monetary Future, March 15, 2016.
But, in my view, this “generous reading of Smith” is really just reading into Smith things that aren’t there. Selgin has actually re-interpreted Smith from ideas in the Theory of Moral Sentiments, and – at the very least – one would have to admit that the explicit theory as presented in The Wealth of Nations needs revision.

Selgin then goes on to argue that Menger, in particular, was well aware of the importance of gift exchange and debt–credit exchanges in premodern human societies before money. Here there are good points. I agree Graeber is unfair to Menger. For example, Menger’s 1892 article “On the Origin of Money” (at least in its English translation) is mercifully free from mathematics, and clearly not filled with “mathematical equations.”

It is true that Menger on this issue and certainly on his other important economic and political views was far from the dogmatic Misesians and Rothbardian anarcho-capitalists who claim his legacy. Menger was a subtle thinker and made real contributions to economics, and was no extremist libertarian.

Here Selgin notes that “Menger understood perfectly well that ‘credit,’ in Graeber’s loose sense of the term, is older than either monetary exchange or barter,” and cites a passage in Menger’s article “Geld” in the Handwörterbuch der Staatswissenschaften (vol. 3), 1892. pp. 730–757. But Selgin cites the 2002 translation of L. B. Yeager and M. Streissler of that article which is actually based on the 3rd edition of 1909, not the original article of 1892.

I have a minor quibble here. I know it is a pedantic point, but did Menger’s views on this evolve over time and move away from his 1890s theory? (see Appendix 1 below).

At any rate, let us put this aside, and move to a much more important point.

If we read Menger’s classical article of 1892 in its English translation by C. A. Foley published in the Economic Journal, we find an interesting qualification that Menger makes to his theory:
It is not impossible for media of exchange, serving as they do the commonweal in the most emphatic sense of the word, to be instituted also by way of legislation, like other social institutions. But this is neither the only, nor the primary mode in which money has taken its origin.” (Menger 1892: 250).
That leaves open the possibility that money can be instituted by a government, though Menger, as far as I can see, clearly did not think that this was the primary and earliest manner by which money had emerged.

But, at the same time, it is also very difficult to see how this isn’t a substantive concession to chartalism by Menger.

That passage by Menger also shows us the divide between Menger’s nuanced view of the origins of money and the stridency of Rothbard:
“[sc. Mises’s] Regression Theorem also shows that money, in any society, can only become established by a market process emerging from barter. Money cannot be established by a social contract, by government imposition, or by artificial schemes proposed by economists.” (Rothbard 2009: 61).
However, Menger remained an advocate of the barter spot trade theory of money’s origins in ancient times, although he was willing to concede what later Austrians have emphatically denied.

But let us return to Selgin’s defence of Menger.

Selgin is essentially arguing the following:
(1) that Menger understood that gift exchange and debt–credit exchanges existed in premodern human societies before money.

(2) that the pure barter economy model from which money emerges is therefore an oversimplification.

(3) that the standard Austrian and neoclassical view on the origins of money can be easily rescued by arguing that it was trade and barter exchange between different communities that was the fundamental origin of money: money emerged in inter-communal barter spot trade as the most saleable good (or goods) became the medium of exchange (as in Menger 1892: 249, though Menger seems to conceive of this happening within communities), and over time money then spread more and more to exchanges within tribes or communities, and internal economies became monetised.
However, I find profound problems with this defence.

For one thing, this is not the mainstream neoclassical view, which does indeed proclaim the naïve theory of the emergence of money from barter spot trade as a universal theory and takes insufficient account of the importance of debt–credit relationships.

Secondly, it is a rather curious state of affairs to see that Selgin is actually proposing a theory that seems very much like the one proposed by Karl Marx in Chapter 2 of volume 1 of Capital:
“Objects in themselves are external to man, and consequently alienable by him. In order that this alienation may be reciprocal, it is only necessary for men, by a tacit understanding, to treat each other as private owners of those alienable objects, and by implication as independent individuals. But such a state of reciprocal independence has no existence in a primitive society based on property in common, whether such a society takes the form of a patriarchal family, an ancient Indian community, or a Peruvian Inca State.

The exchange of commodities, therefore, first begins on the boundaries of such communities, at their points of contact with other similar communities, or with members of the latter. So soon, however, as products once become commodities in the external relations of a community, they also, by reaction, become so in its internal intercourse.
The proportions in which they are exchangeable are at first quite a matter of chance. What makes them exchangeable is the mutual desire of their owners to alienate them. Meantime the need for foreign objects of utility gradually establishes itself. The constant repetition of exchange makes it a normal social act. In the course of time, therefore, some portion at least of the products of labour must be produced with a special view to exchange. ....

The necessity for a value-form grows with the increasing number and variety of the commodities exchanged. The problem and the means of solution arise simultaneously. Commodity-owners never equate their own commodities to those of others, and exchange them on a large scale, without different kinds of commodities belonging to different owners being exchangeable for, and equated as values to, one and the same special article. Such last-mentioned article, by becoming the equivalent of various other commodities, acquires at once, though within narrow limits, the character of a general social equivalent. This character comes and goes with the momentary social acts that called it into life. In turns and transiently it attaches itself first to this and then to that commodity. But with the development of exchange it fixes itself firmly and exclusively to particular sorts of commodities, and becomes crystallised by assuming the money-form. The particular kind of commodity to which it sticks is at first a matter of accident. Nevertheless there are two circumstances whose influence is decisive. The money-form attaches itself either to the most important articles of exchange from outside, and these in fact are primitive and natural forms in which the exchange-value of home products finds expression; or else it attaches itself to the object of utility that forms, like cattle, the chief portion of indigenous alienable wealth. Nomad races are the first to develop the money-form, because all their worldly goods consist of movable objects and are therefore directly alienable; and because their mode of life, by continually bringing them into contact with foreign communities, solicits the exchange of products.” (Marx 1906: 99–101).

“An adequate form of manifestation of value, a fit embodiment of abstract, undifferentiated, and therefore equal human labour, that material alone can be whose every sample exhibits the same uniform qualities. On the other hand, since the difference between the magnitudes of value is purely quantitative, the money commodity must be susceptible of merely quantitative differences, must therefore be divisible at will, and equally capable of being re-united. Gold and silver possess these properties by nature.” (Marx 1906: 102).
Of course, it is not exactly the same. But, if we just ditch Marx’s mystical labour theory of value nonsense, we actually have a theory here not far different from the one Selgin proposes.

According to Marx, money emerges by necessity from barter exchange of commodities between different communities (Marx 1990: 181–182). Money must be a uniform, portable, fungible and divisible commodity. It is implied that money then spread into communities and it monetised exchanges within tribes or communities.

Something like this was probably also the view proposed by the German Historical School economists of the 19th and early 20th centuries such as Max Weber (1978: 673–674) and Karl Bücher (1901), who argued that money emerged from barter between different societies, not within societies (Karl Polanyi may also have held a position close to this).

But, once again, even if we adopt this revised theory of money’s origins, modern anthropology and history suggest that it has serious problems and is in need of revision.

Our starting point is not – repeat not – that there is nothing of value in Smith’s or Menger’s theories, but (1) that it is not a universally applicable theory, and (2) that there are other important ways by which money can emerge.

Even Graeber does not deny that money in some historical circumstances can emerge from barter between strangers, especially in long distance trade (Graeber 2011: 75).

Now Selgin already admits that primitive money-less societies are frequently dominated by debt/credit transactions, or “gift exchange,” not by barter spot trades.

I assume he also seems to accept that even in cases where goods exchange for goods in spot trades, social relations can complicate matters considerably, and historically barter seems to have been prevalent between one community and another, or, that is to say, between people who were strangers and where relationships were implicitly or explicitly hostile (as in Graeber 2011: 29–30).

While a non-enumerated system of debts/credits or gift exchange might not give rise to money, there is clearly a role for debt in the history of money (Graeber 2011: 40). In the real world, gift exchange and debt/credit arrangements existed long before money, and societies could develop an effective system of exchange in which debt/credit or gift exchange transactions were the predominant system (Graeber 2011: 40).

But, under the theory that Selgin proposes, doesn’t this mean that societies that had no interest in large-scale trade with foreigners had little reason to develop money? Wouldn’t it also imply that the double coincidence of wants problem would be largely overcome in such an early tribal or isolated society?

More likely, such communities would sometimes develop what anthropologists call “non-commercial money” or “ceremonial money,” which is non-commercial in the sense that it is not used for everyday purchases of goods and services, or only rarely for such ordinary goods. It is thus non-commercial in the sense that it is not a universal medium of exchange. The purpose of such non-commercial money is social (see also here).

Graeber calls non-commercial money “primitive monies” and gives examples such as the shell money in the Americas or Papua New Guinea, cattle money in Africa, bead money, feather money, and so on. These are rarely used to buy everyday items in the societies that use them. Instead, they are employed in social relations like marriages and to settle disputes (Graeber 2011: 60).

At this point, we should consider the views of older anthropologists. In 1949, A. H. Quiggin published A Survey of Primitive Money: The Beginnings of Currency (London).

The summary of the author’s views on the origin of money is worth quoting:
“Writers on the origin of the use of money often start with a consideration of barter and its inconveniences. From ‘silent trade’ (a primitive though abnormal form of barter) they trace the evolution of trading and money side by side, relying mainly on literary evidence for probing into the past.

This study relies mainly on the tangible evidence of the actual types of primitive money or money-substitutes used by ‘unrisen’ people and others all over the world, and is concerned with the purposes, when discoverable, for which they were used. The evidence suggests that barter in its usual sense of exchange of commodities was not the main factor in the evolution of money. The objects commonly exchanged in barter do not develop naturally into money and the more important objects used as money seldom appear in ordinary everyday barter. Moreover, the inconveniences of barter do not disturb simple societies. The variety of material and the complexities of uncivilized attitudes towards money preclude generalizations, but the evidence appears to support the following line of argument.

In the beginning Man lived in self-supporting and self-contained groups. Except in an area where provisions are unlimited, a society depending on hunting and food-gathering for its subsistence is necessarily unsociable and ‘has no truck’ with its neighbours. Early exchanges were in the way of present-giving, and were expressions of friendship with no ulterior economic purpose, although the latter – an expectation of an adequate or even improved return – cannot be excluded from human dealings. Present-giving or gift-exchange, seen in simple forms in the Andamans, Torres Straits or New Zealand, may develop into elaborate ceremonial as in Fiji or the North-West of America, but remain distinct from trading with money.

Barter develops between areas of contrasted produce, such as coastal and inland, forested and open country. We see the barter of fish or shells for vegetables, game for bananas, &c., in Melanesia or the Congo, and the establishment of regular markets. Trading voyages such as those of Torres Straits and New Guinea take us a stage further by the introduction of conventional presents. But so far there is no need for any medium of exchange such as is commonly described as money.

This is the state of affairs over about half the world at the present day. Barter suffices for most of the natives of Australia, New Zealand and the islands of the Pacific, and for the less-advanced peoples of Africa, Asia and the Americas, where native economy is not upset by the trader and the missionary.

The use of a conventional medium of exchange, originally ‘full-bodied’ but developing into ‘token’ money, is first noted in the almost universal customs of ‘bride-price’ and wergeld. When sister-exchange is not practicable, some other value must be substituted; where life for life is not demanded, some equivalent must be found. The history of ‘bride-price’ and wergeld (which has yet to be written) shows how formal the customary gifts become, fitted to definite scales of value. It is not without significance that in any collection of primitive currency the majority of the items are described as ‘used in bride-price’.

When once a system of conventional gifts or payments with a definite scale of values has been established (and this is necessary for ‘bride-price’ and for wergeld) the first steps are taken in the evolution of money. It develops thereafter in response to human needs into the accepted medium of exchange. Nutzgeld [viz. “useful money” or useful objects in exchange – LK] still remains Nutzgeld. Cattle may constitute wealth and form a standard of value. They cannot, strictly speaking, be called money. Money, to be generally acceptable, needs more convenient material and finds the four essential qualities (portable, divisible, durable, recognizable) in shells, beads or metals. Two further qualities have been shown to be necessary, one geographical and one more difficult to define.

The objects that come to be used as money are mainly non-local, or if local are the product of a special area or a special class; and they have prestige or essential virtue, religious or magical. Cowries and beads, most universal of all forms of primitive money, have magical as well as monetary value and still hold their own over a large part of the world, though everywhere disappearing now with the advent of the trader and trade tobacco. Metals best illustrate the transition from ‘full-bodied’ to ‘token-’ money. The spears and hoes of Africa, the knives and spades of China, and the spits of Argos are familiar examples. The tools may become amorphous and valued according to their weight in metal, or survive as attenuated imitations of their former selves. Metal, whether gold, silver, copper, iron or tin, is everywhere useful and everywhere valued, and estimated by size, shape or weight. Ingots are preliminary stepping-stones to coins. Ingots, as lumps or bars, develop in response to local needs or whims in special forms, such as manillas, Katanga crosses and Kissi pennies, Malay hats and Siamese bullets, or our own currency bars and ‘ring-money’.

To us, looking backward, the next step appears obvious and inevitable, but it was only in rare spots (possibly only in one rare spot) in the Old World that the final stage was reached, and definite weights of metal, rounded, flattened and stamped, can be called coins. Here the study of primitive money comes to an end.” (Quiggin 1949: 321–322).
It is depressing really how good empirical research like this is often forgotten and ignored in neoclassical economics, and there are two fundamental points to be taken from Quiggin’s analysis:
(1) many primitive people were able to exist without the modern developed form of money, and even where it arose in trade between foreigners it seems to have been largely limited to that sphere.

(2) the origin of money within a society can also be linked to social customs like bride-price and wergild (“man-money” or compensation for murder).
These conclusions seem to have been widely held by early 20th century anthropologists. Paul Einzig, for example, using the anthropological work of his day, argued a long time ago that “money first developed to serve matrimonial, political or religious payments was only later adopted gradually for commercial purposes” (Einzig 1948: 984).

So let us now focus on (2). I rely on Philip Grierson’s The Origins of Money (1977) (see here on that book).

As we have seen, in primitive societies, there often arises non-commercial money or ceremonial money. It is mainly used in social interactions, often formal social events such as marriage, wergild and bloodwealth payments, political relations (e.g., potlatch, moka), and fines and compensations (compensation for adultery, or for things lost), and may only be rarely used, if at all, for everyday purchases or commercial transactions (Grierson 1977: 15–16).

Sometimes this non-commercial money develops into money as a more general medium of exchange in some societies (though often enough it remains in its traditional role in others where no universal medium of exchange arises). So where does the origin of non-commercial money as a general medium of exchange and unit of account come from?

Grierson (1977: 19) proposes that the social custom of wergeld and wergeld-like customs are the answer. Wergeld (literally, “man-money”) is the paying of compensation for murder or other injuries and even theft of personal property. The object of wergild is to stop blood feuds and violence in revenge, and to provide an adequate measure of the things lost, as well as compensation. The objects that arose as standards of value as non-commercial money in tribal wergild payments did not necessarily arise by barter spot trade of the most saleable commodity (Grierson 1977: 21; 28–29).

Often in tribal societies objects of high social status or conferring “prestige” or even thought to have magical power will function as non-commercial money. Thus such non-commercial money doesn’t necessarily arise by barter spot trade as the most saleable commodity.

Most interesting is the linguistic evidence from many societies which shows how the word for money arose etymologically from concepts related to wergild and debt. The English word “pay” comes via French payer from the Latin word pacare, meaning “to pacify,” “make peace with.” In certain societies, non-commercial money arose as a standard for measuring value related to wergild-like customs and possibly even things like bride-wealth, but did not necessarily develop into general purpose money/commercial money.

Compensation payments are made in various goods, such as cattle, bondmaids, and precious metal, but it is likely a common unit of account was developed to simplify calculation of payments, which later spread to the wider community in economic transactions.

Where commercial money arose from non-commercial money, Grierson makes the following argument:
“… where societies have developed the notion of money as a general measure of value, it will, I believe, most often be found that a system of legal compensation for personal injuries, at once inviting mutual comparison and affecting every member of the community, lay behind them.” (Grierson 1977: 29).
So it is possible that in some societies commercial money arose from its previous role in systems of legal compensation.

But this clearly isn’t the end of the story either. What about ancient advanced civilisations?

The origin of money in ancient Mesopotamia appears to be in the development of an abstract money of account in the temple and palace institutions. These temples and palaces were institutions with large internal centrally planned economies, with complex weights and measurements for internal accounting of the products produced, received and distributed, and rent and interest owed. Many prices were set and administered in the money of account which developed from weight units. The two units of account were (1) the shekel of silver (which was equal to the monthly grain ration) and (2) barley (Hudson 2004). Silver money of account spread to the private economy mostly as a means of reckoning debts to temples and palaces (Hudson 2004: 115). But many ordinary people could pay in commodities, and the administered pricing system in terms of silver/grain that was developed in the temples was to assist in calculation of payments in kind.

In ancient Egypt, money appears as the most important unit of account called the deben (or uten), which was a unit of weight, originally equated to 92 (or 91) grams (Henry 2004: 92; there was also the unit called the khar for measuring wheat or barley, and 1 khar was equivalent to 2 deben of bronze). The measure of value for various goods was thus fixed weight units and historically no doubt these units arose from copper, silver, grain and gold. The unit of account system appears to have been developed by complex palace, government and temple institutions for internal accounting. While goods came to be denominated in terms of deben, in early times during the period of the Old Kingdom there were no physical deben changing hands in the private economy. That is to say, the deben did not function as a physical means of payment, and did not emerge by barter spot transactions as the most saleable medium of exchange. Even though goods and services were measured in a deben unit of account, payment was made in goods.

An important element in both these historical processes was the institution (or institutions) where surplus products were stored from taxation, tribute and gifts. These institutions dealt with complex flows, in and out, of goods: they were palace and temple complexes. Accounting systems, weight measures and writing are connected with just such institutions, and, importantly, some abstract unit of account arose by which to measure relative values of stored goods and inflows or outflows of goods. Since loans were also no doubt made from surplus products stored, repayment of loans in kind was facilitated by a unit of account.

In primitive human societies by the end of the Stone Age (c. 2.9 million years to 4500 BC), before the literate urban civilizations, agricultural communities developed where surpluses were stored, most probably held as a communal resource. The question of how the origins of a unit of account or measure of value could be related to the emergence of stored communal wealth is an interesting research question that deserves further study.

It is quite possible that silver and gold were used as non-commercial money or ceremonial money in ancient societies too before they became abstract units of account.

So even in developed ancient civilisations with cities we have huge collectivist temples and governments (sometimes presided over by god-kings) with large-scale economic and central planning.

So where are they in neoclassical theory and Selgin’s theory?

Given its very scarcity, it is unlikely that silver would have arisen as a unit of account and medium of exchange in, say, Mesopotamia from internal barter trade as the most saleable good precisely because there wasn’t enough of it.

More likely, it was imposed as a unit of account from above, by the collectivist temples and governments. The ancient government–temple complex may well have adopted it because it was important in foreign and international trade, but even here that requires yet more massive revision to Selgin’s theory, taking account of ancient institutions.

But it may well have been adopted merely because it was high-prestige object with a religious significance and from its role as a weight unit. That is, it arose partly because as a weight unit it was easily convertible into an abstract unit of account.

We could also examine the origin of money in ancient Greece, which I have done in detail in this post.

In short, before coins the state of affairs was this:
(1) cattle or oxen functioned as a largely abstract unit of account (but not a common medium of exchange) and

(2) iron spits might have been a very limited or weak medium of exchange.
In Homer’s epics which reflect the social reality of Dark Age and early Archaic period Greece, cattle or oxen are a type of unit of account, but the actual means of payment tend to be many other types of goods, not just cattle (Peacock 2011: 49–54).

Now one might argue that cattle did become the most common medium of exchange but then receded in importance to become a mere unit of account, but there are serious problems with this view. First, the emergence of a “cattle/ox” unit of account in Greece appears to be related to religion and cult offerings, not emergence of cattle as the most saleable good.

While cattle no doubt had value in market trades, they were an important sacrificial animal and offering to the gods. The Greeks appear to have developed a cattle or ox unit of account derived from the value these animals had in ritual and sacrifice (Seaford 2004: 61; the view was originally proposed by Laum 1924). Religious rituals and then temples had a preeminent place in ancient Greek society, and the city government’s major responsibility was to honour, appease and placate the gods by offering sacrifices. In this sense, the ancient Greek temple and city are not separate entities, but really one and the same. So the emergence of an ox unit of account can be seen as another state-based, institutional process affecting economic life.

Metal coinage in Lydia and Greece was an invention of the state, and the first Lydian coinage was struck in electrum and used to pay soldiers and mercenaries. This was most probably a high prestige object and perhaps even non-commercial money. At most, it was simply one of many goods used in conventional barter trades: there is no convincing evidence that it was the reigning medium of exchange (money) that had already emerged as the most saleable good in spot barter trades before it was adopted by the Lydian state.

Instead, electrum was a high prestige commodity selected by the state, standardised and used as a form of payment as wages. Its subsequent rise in market trades on a significant scale as a common medium of exchange was then induced by the exchange of these coins for goods by soldiers, and the need to acquire the coins themselves to pay taxes.

Again, for full analysis of the emergence of money in Greece, see my post here.

So, finally, we have a vast, vast body of empirical evidence that the orthodox Mengerian and neoclassical explanations of the origins of money are deeply flawed and not adequate to describe the full complexity of human history. I am afraid even Selgin’s revised Mengerian theory, curiously similar to that of Karl Marx, suffers from the same limitations.

Appendix 1: Menger’s Writings on the Nature and Origin of Money
The writings of the Austrian economist Carl Menger on money extend well behind his classic article of 1892, and include the following:
Menger, C. 2007. Principles of Economics (trans. Grundsätze der Volkswirtschaftslehre [1st edn. 1871] by J. Dingwall and B. F. Hoselitz), Ludwig von Mises Institute, Auburn, Alabama. pp. 257–285.

Menger, C. 1892. “Geld,” in Handwörterbuch der Staatswissenschaften (vol. 3). 730–757.

Menger, C. 1892. “On the Origin of Money” (trans. C. A. Foley), Economic Journal 2: 238–255.

Menger, C. 1909. “Geld,” in J. Conrad et al. (eds.), Handwörterbuch der Staatswissenschaften (vol. 4; 3rd edn.), Fischer, Jena. 555–610.

Menger, C. 1923. Grundsätze der Volkswirtschaftslehre (2nd rev. edn.), Hölder-Pichler-Tempsky, Vienna.

Menger, C. 2002 [1909]. “Money” (trans. L. B. Yeager and M. Streissler), in M. Latzer and S. W. Schmitz (eds.), Carl Menger and the Evolution of Payments Systems, Edward Elgar, Cheltenham, UK. 25–108. [N.B. this is a translation of Menger 1909.].
We must remember that Menger’s treatment of money developed in the course of his life.

Now L. B. Yeager and M. Streissler’s translation of Menger’s article on money from the Handwörterbuch der Staatswissenschaften is based on the 3rd edition of 1909, not the original article of 1892.

Was the passage that Selgin cites from Menger (2002 [1909]) actually in Menger’s 1892 article “Geld”? This is a pedantic point, but an interesting one, since Menger may have developed and modified this 1890s views as he got older.

Further Reading
“The Nature, Origin and History of Money 101.”

“A Note on Menger on the Nature and Origin of Money,” July 28, 2012.

“Menger on the Origin of Money,” January 5, 2012.

“Menger’s Nuanced View on the Origin of Money,” November 6, 2012.

“Observations on Non-Commercial Money,” February 18, 2012.

“Philip Grierson on the Origin of Money,” March 21, 2012.

“The Origin of Money and Coinage in Western Civilisation: The Case of Ancient Greece,” April 5, 2013.

Bücher, K. 1901. Industrial Evolution (trans. S. Morley Wickett), H. Holt and Co., New York.

Einzig, Paul. 1948. “New Light on the Origin of Money,” Nature 162.4130 (25 December): 983–985.

Graeber, D. 2011. Debt: The First 5,000 Years. Melville House, Brooklyn, N.Y.

Grierson, P. 1977. The Origins of Money. Athlone Press and University of London, London.

Henry, J. F. 2004. “The Social Origins of Money: The Case of Egypt,” in L. R. Wray (ed.), Credit and State Theories of Money: The Contributions of A. Mitchell Innes, Edward Elgar, Cheltenham, UK. 79–98.

Hudson, M. 2004. “The Archaeology of Money: Debt Versus Barter Theories of Money’s Origins,” in L. R. Wray (ed.), Credit and State Theories of Money: the Contributions of A. Mitchell Innes, Edward Elgar, Cheltenham. 99–127.

Laum, B. 1924. Heiliges Geld: eine historische Untersuchung über den sakralen Ursprung des Geldes. Mohr, Tübingen.

Marx, Karl. 1906. Capital. A Critique of Political Economy (vol. 1; rev. trans. by Ernest Untermann from 4th German edn.). The Modern Library, New York.

Marx, Karl. 1990. Capital. A Critique of Political Economy. Volume One (trans. Ben Fowkes). Penguin Books, London.

Menger, C. 1892. “On the Origin of Money,” Economic Journal 2: 238–255.

Peacock, M. S. 2011. “The Political Economy of Homeric Society and the Origins of Money,” Contributions to Political Economy 30: 47–65.

Quiggin, A. H. 1949. A Survey of Primitive Money: The Beginnings of Currency. Methuen, London.

Rothbard, M. N. 2009. The Essential von Mises. Ludwig von Mises Institute, Auburn, Alabama.

Seaford, R. 2004. Money and the Early Greek Mind: Homer, Philosophy, Tragedy. Cambridge University Press, Cambridge.

Selgin, G. “Those Dishonest Goldsmiths,” revised January 20, 2011

Weber, M. 1978. Economy and Society: An Outline of Interpretive Sociology (eds. G. Roth and C. Wittich; trans. E. Fischoff et al.), University of California Press, Berkeley and London.

Tuesday, March 29, 2016

Human Consciousness, Artificial Intelligence and Popper’s Three World Ontology

Long-time commentator Ken B has drawn my attention to this post by Gene Callahan that complains that the Turing Test doesn’t really test for real intelligence.

In essence, I actually agree on this point, though I would reject Callahan’s dualism.

The endless debates about whether machines or software have real intelligence generally tend to suffer from a shoddy fallacy of equivocation. Generally, people who want to defend the real intelligence of machines make arguments like this:
(1) machines have intelligence and can think.

(2) a Turing test can show whether a machine has intelligence and can think.

(3) the brain is just an information processing machine, and so since computers are also information processing machines, a sufficiently complex computer should eventually have intelligence.
Yes, but what do you mean by “intelligence” and “thinking” in (1)?

Do you mean the conscious intelligence and thinking that I am experiencing right now? Actually, there is no good reason to think software can think in this way, because (3) is unproven and a non sequitur.

It does not follow that the consciousness of the human mind is just information processing that can be reproduced in other synthetic materials, such as in silicon chips in digital computers. After all, DNA and its behaviour exhibit a type of information processing and storage as well, but DNA is not conscious.

The conscious thinking intelligence that humans have is most probably a higher-level emergent property from specific biological processes in the brain.

As the analytic philosopher John Searle points out in the video below, all the empirical evidence suggests that consciousness is a biological phenomenon in the brain causally dependent on neuronal processes and biochemical activity, but one that can be explained by physicalist science, not some discredited supernatural ideas about souls or Cartesian dualism (see also Searle 1990; Searle 2002; Searle 1992).

Now the crude behaviourist Turing test does not test for whether an entity has conscious life, but merely whether it simulates human intelligence. Thus even if future computers all start passing Turing tests, it is not going to be some shocking milestone in human history: all it will show is that software programs have become sophisticated enough to fool us into thinking machines have conscious minds as we do, even though they do not.

Searle’s famous Chinese room argument is a devastating blow to the cult of AI.

But there are some additional philosophical points here. There is nothing supernatural about saying that, physically speaking, you need brain chemistry to produce a conscious intelligent mind.

And, philosophically speaking, the key to cracking the problem of human conscious intelligence is Karl Popper’s pluralist ontology. This is not a supernaturalist ontology, but naturalistic.

The key to understanding complexity in our universe is this: emergent properties. Karl Popper’s Three World Ontology gives us the best conceptual framework in which to understand our universe and the human mind.

Now this is where Searle goes philosophically wrong: his adherence to a crude monist materialism. Searle is unwilling to accept that (1) some emergent properties – like the human mind – are so remarkable that they make a new ontological category, and (2), because of (1), we have good reason to consider Popper’s three worlds ontology as a realistic metaphysics.

So what is Karl Popper’s pluralist ontology?

Karl Popper’s Three Worlds system is a pluralist ontology that classifies all objects in our universe, even though it is quite different from dualism. Ultimately, all objects in Popper’s world are causally dependent on matter and energy and their relations and permutations, so that – despite being ontologically pluralist – it has a fundamentally naturalist and physicalist/materialist basis.

We can set out the Three World Ontology in the following diagram.

World 1 give rises to World 2 objects, and World 2 in turn gives rise to World 3 objects, World 2 and World 3 being emergent properties from the each lower-level world respectively.

These three worlds can be described as follows:
World 1: the fundamental world of spatiotemporal physical objects and events, matter and energy and even biological systems considered merely as only complex physical systems. It further divided into a fundamental distinction between:
(1) the world of quantum mechanics, and

(2) the macroscopic world of Newtonian physics and Einstein’s special and general theories of relativity.
In some manner that is not properly understood our macroscopic world is an emergent property from the world of quantum mechanics.

World 2: the world of conscious intelligent human minds with intentionality and mental states (and, for example, any alien minds as complex as ours that have evolved on other planets, if they exist). True conscious intelligence is a property of minds in this world.

World 3: both
(1) unembodied objects such as abstract entities (and, most probably, universals), mathematical entities, social institutions and objective knowledge, and
(2) embodied entities that are products of human design, engineering and production such as buildings, books, clothes, money, and art works considered as human objects.
We must remember that “embodied entities” are members of both World 3 and World 1, but in different senses: e.g., considered purely physically as spatiotemporal objects describable in scientific terms, they belong to World 1; but when considered as human cultural artefacts, they belong to World 3.
The conscious intelligent human mind is an emergent property from World 1 brain states, but is so powerfully unique it requires a new ontological category.

A true artificial intelligence that would have the same type of conscious intelligent mind as the human mind will have to directly reproduce or replicate the biological processes in the brain that cause consciousness. An “artificial” intelligence – in the sense of not being a normal human being – will probably need to have organic or biochemical structures in its “brain” in order for it to be fully and truly conscious.

Such entities, if they were fully conscious, would create all sorts of ethical issues. They would probably have to be imbued with moral/ethical principles as humans are, for example. Probably they would have to be granted some kind of human rights at some point, so that we could not treat them as slaves. And what would they do? What work would they perform?

The whole notion that your desktop computer could become as conscious as you are should be a horrifying thought. Wouldn’t it effectively be a slave?

At any rate, the whole issue of whether truly conscious artificial intelligence can be created is a matter for a future science that has first completely mastered what human (and higher animal) consciousness actually is.

To return to the issue of ontology, Popper summed up his system as follows:
“To sum up, we arrive at the following picture of the universe.

There is the physical universe, world 1, with its most important sub-universe, that of the living organisms.

World 2, the world of conscious experience, emerges as an evolutionary product from the world of organisms.

World 3, the world of the products of the human mind, emerges as an evolutionary product from world 2.

In each of these cases, the emerging product has a tremendous feedback effect upon the world from which it emerged. For example, the physico-chemical composition of our atmosphere which contains so much oxygen is a product of life – a feedback effect of the life of plants. And, especially, the emergence of world 3 has a tremendous feedback effect upon world 2 and, through its intervention, upon world 1.

The feedback effect between world 3 and world 2 is of particular importance. Our minds are the creators of world 3; but world 3 in its turn not only informs our minds, but largely creates them. The very idea of a self depends on world 3 theories, especially upon a theory of time which underlies the identity of the self, the self of yesterday, of today, and of tomorrow. The learning of a language, which is a world 3 object, is itself partly a creative act and partly a feedback effect; and the full consciousness of self is anchored in our human language.

Our relationship to our work is a feedback relationship: our work grows through us, and we grow through our work.

This growth, this self-transcendence, has a rational side and a non-rational side. The creation of new ideas, of new theories, is partly non-rational. It is a matter of what is called ‘intuition’ or ‘imagination’. But intuition is fallible, as is everything human. Intuition must be controlled through rational criticism, which is the most important product of human language. This control through criticism is the rational aspect of the growth of knowledge and of our personal growth. It is one of the three most important things that make us human. The other two are compassion, and the consciousness of our fallibility.” (Popper 1978: 166–167).
At any rate, Popper’s three world ontology has the following virtues:
(1) it avoids the fallacy of strong reductionism;

(2) it is compatible with a moderate realist or conceptualist view of abstract entities and universals;

(3) it understands that human minds are an important, real entity in the universe and is consistent with John Searle’s biological naturalist theory of the mind, even though minds are a type of emergent physical property (for why Searle is not a property dualist, see Searle 2002);

(4) it is compatible with the finding of modern science that emergent properties are a fundamental phenomenon in our universe;

(5) it is compatible with downwards causation;

(6) it is ultimately consistent with the view that higher-level worlds are causally dependent for their existence on lower-level worlds.
Jesper Jespersen (2009) uses Popper’s three worlds ontology in his Critical Realist methodology for Post Keynesian economics, and it is highly relevant for the ontological and epistemological basis of economics.

Further Reading
“More on Karl Popper’s Three World Ontology,” September 6, 2013.

“Karl Popper’s Three World Ontology,” September 4, 2013.

“Limits of Artificial Intelligence,” February 24, 2014.

Jespersen, Jesper. 2009. Macroeconomic Methodology: A Post-Keynesian Perspective. Edward Elgar, Cheltenham, Glos. and Northampton, MA.

Popper, Karl R. 1978. “Three Worlds,” The Tanner Lecture on Human Values, Delivered at the University of Michigan, April 7, 1978.‎

Searle, John. R. 1980. “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3: 417–424.

Searle, John. R. 1980. “Intrinsic Intentionality,” Behavioral and Brain Sciences 3: 450–456.

Searle, John R. 1982. “The Chinese Room Revisited,” The Behavioral and Brain Sciences 5: 345–348.

Searle, John. R. 1990. “Is the Brain a Digital Computer?,” Proceedings and Addresses of the American Philosophical Association 64: 21–37.

Searle, John R. 1992. The Rediscovery of the Mind. MIT Press, Cambridge, Mass and London.

Searle, J. 2002. “Why I Am Not a Property Dualist,” Journal of Consciousness Studies 9.12: 57–64.

Monday, March 28, 2016

Trump’s Wall versus Spain’s Wall: The Left Wing Hysteria about Trump’s Proposed Border Wall

There is a country called Spain. Many people probably do not know that Spain still has a few small strips of territory in North Africa called the “plazas de soberanía” (“places of sovereignty”), such as Ceuta and Melilla. I am not sure how attached the people of Spain are to these outposts in North Africa, but that is not the issue here.

In some of these places, the Spanish government has built walls or fences to protect its territories.

Here is a photo of the Melilla border fence between Morocco and Spain’s Melilla territory.

The law is the law. It is illegal for people to just immigrate into a different nation. Spain is neither fascist nor authoritarian nor racist for having its wall. It’s a plain common sense way to solve the problem of illegal immigration.

America’s problem with illegal immigration is far, far worse than Spain’s. It seems to be plain common sense that if America cannot police its border with Mexico, then some kind of wall is probably the only solution to the problem.

Yet people on the left are screaming hysterical abuse at the very idea. Apparently it’s racist; it’s fascist; it’s Nazism reborn.

But it’s none of these things. In reality, even Bernie Sanders knows that open borders and mass illegal immigration are terrible threats to America’s working class and real wages.

He’s well aware of the problem, as we see in the video below.

Unfortunately, it is unlikely that Bernie could ever really do anything effective to stop mass illegal immigration. Why? Because Bernie is too kind and too compassionate a man. And I mean that as a genuine compliment to Bernie Sanders, because I like him a lot.

Unfortunately, to really do something effective to stop open borders, Bernie – even as a democratic socialist – would have to build a wall or fence on America’s border with Mexico too, rather like Spain’s fences to protect its North African territories. This is a hard reality of life.

Left-wing people will do nothing but discredit themselves by ignoring reality.

Marx’s Capital, Volume 1, Chapter 21: A Critical Summary

Chapter 21 of volume 1 of Capital is called “Piece Wages” and deals with wages paid by the number of output goods produced.

A “piece wage” is a wage payment by quantity of products or units produced by a worker.

Time wages and piece wages both exist and sometimes side by side. Marx argues that “[w]ages by the piece are nothing else than a converted form of wages by time, just as wages by time are a converted form of the value or price of labour-power” (Marx 1906: 602).

But piece wages conceal surplus labour value and unpaid labour time:
“Piece-wages do not, in fact, distinctly express any relation of value. It is not, therefore, a question of measuring the value of the piece by the working time incorporated in it, but on the contrary of measuring the working-time the labourer has expended, by the number of pieces he has produced. In time-wages the labour is measured by its immediate duration, in piece-wages by the quantity of products in which the labour has embodied itself during a given time. The price of labour-time itself is finally determined by the equation; value of a day's labour = daily value of labour-power. Piece-wage is, therefore, only a modified form of time-wage.

Let us now consider a little more closely the characteristic peculiarities of piece-wages.

The quality of the labour is here controlled by the work itself, which must be of average perfection if the piece-price is to be paid in full. Piece-wages become, from this point of view, the most fruitful source of reductions of wages and capitalistic cheating.

They furnish to the capitalist an exact measure for the intensity of labour. Only the working-time which is embodied in a quantum of commodities determined beforehand and experimentally fixed, counts as socially necessary working time, and is paid as such. In the larger workshops of the London tailors, therefore, a certain piece of work, a waistcoat e.g. is called an hour, or half an hour, the hour at 6d. By practise it is known how much is the average product of one hour. With new fashions, repairs, etc., a contest arises between master and labourer, whether a particular piece of work is one hour, and so on, until here also experience decides. Similarly in the London furniture workshops, etc. If the labourer does not possess the average capacity, if he cannot in consequence supply a certain minimum of work per day, he is dismissed.” (Marx 1906: 604–605).
So piece wages can control and ensure the intensity and quality of labour, and supervision of workers becomes to a great degree unnecessary (Marx 1990: 695).

Piece wages therefore become the basis of domestic labour, sweat shop labour and subletting or subcontracting of labour, and parasitical capitalist middlemen arise in the process (Marx 1990: 695). Piece wages therefore produce a hierarchy of wages (Marx 1990: 695).

As Marx says:
“Given piece-wage, it is naturally the personal interest of the labourer to strain his labour-power as intensely as possible; this enables the capitalist to raise more easily the normal degree of intensity of labour. It is moreover now the personal interest of the labourer to lengthen the working day, since with it his daily or weekly wages rise. This gradually brings on a reaction like that already described in time-wages, without reckoning that the prolongation of the working day, even if the piece-wage remains constant includes of necessity a fall in the price of the labour.” (Marx 1906: 606–607).
Marx even thinks that the “piece-wage is the form of wages most in harmony with the capitalist mode of production” (Marx 1906: 608).

As the productivity per hour of the piece wage worker increases with new technology, so the actual piece wage must fall, given the falling amount of socially necessary labour embodied in each individual output product (Marx 1990: 699).

So Marx argues that employers keep the piece wage rates at the subsistence level to continue to steal surplus labour-time, and the wage per piece tends to be the daily value of the maintenance and reproduction of labour divided by the average number of pieces produced per day by the average worker (Brewer 1984: 65).

Brewer, Anthony. 1984. A Guide to Marx’s Capital. Cambridge University Press, Cambridge.

Harvey, David. 2010. A Companion to Marx’s Capital. Verso, London and New York.

Marx, Karl. 1906. Capital. A Critique of Political Economy (vol. 1; rev. trans. by Ernest Untermann from 4th German edn.). The Modern Library, New York.

Marx, Karl. 1990. Capital. A Critique of Political Economy. Volume One (trans. Ben Fowkes). Penguin Books, London.

Sunday, March 27, 2016

Marx on Wage Determination and the Classical Economists

In Chapter 19 of volume 1 of Capital Marx quotes the Classical economics on wage determination:
“Classical political economy borrowed from every-day life the category ‘price of labour’ without further criticism, and then simply asked the question, how is this price determined? It soon recognized that the change in the relations of demand and supply explained in regard to the price of labour, as of all other commodities, nothing except its changes, i.e., the oscillations of the market price above or below a certain mean. If demand and supply balance, the oscillation of prices ceases, all other conditions remaining the same. But then demand and supply also cease to explain anything. The price of labour, at the moment when demand and supply are in equilibrium, is its natural price, determined independently of the relation of demand and supply. And how this price is determined, is just the question. Or a larger period of oscillations in the market-price is taken, e.g., a year, and they are found to cancel one the other, leaving a mean average quantity, a relatively constant magnitude. This had naturally to be determined otherwise than by its own compensating variations. This price which always finally predominates over the accidental market-prices of labour and regulates them, this ‘necessary price’ (physiocrats) or ‘natural price’ of labour (Adam Smith) can, as with all other commodities, be nothing else than its value expressed in money. In this way political economy expected to penetrate athwart the accidental prices of labour, to the value of labour. As with other commodities, this value was determined by the cost of production. But what is the cost of production—of the labourer, i.e., the cost of producing or reproducing the labourer himself? This question unconsciously substituted itself in political economy for the original one; for the search after the cost of production of labour as such turned in a circle and never left the spot. What economists therefore call value of labour, is in fact the value of labour-power, as it exists in the personality of the labourer, which is as different from its function, labour, as a machine is from the work it performs. Occupied with the difference between the market-price of labour and its-so-called value, with the relation of this value to the rate of profit, and to the values of the commodities produced by means of labour, &c., they never discovered that the course of the analysis had led not only from the market prices of labour to its presumed value, but had led to the resolution of this value of labour itself into the value of labour-power.” (Marx 1906: 589–590).
Marx essentially endorses in this chapter the view that wages fluctuate above and below an equilibrium value: the “necessary price” or “natural price” of labour.

But what Marx does say is that the Classical economists did not really understand the real nature of the “natural price”: they did not understand that this was the value of the maintenance and reproduction of labour-power, a subsistence wage.

As we know from other passages, Marx also rejected the Malthusian population theory, but even so the important point remains: Marx sees capitalism as driving wages towards an equilibrium subsistence wage, the value of the maintenance and reproduction of labour-power (sometimes with a moral and historical element).

Supply and demand might cause wage rates to temporarily go above or below this level, but the tendency of capitalism is to drive the real wage back to the subsistence level.

There are powerful forces driving the real wage back to the subsistence level as follows:
(1) capitalists try to reduce the real wage to and even below subsidence level;

(2) capitalists reduce the price of basic commodities required for subsistence and so reduce the necessary part of the working day and hence the value of the maintenance and reproduction of labour (see Chapter 12 of volume 1 of Capital);

(3) continuous technological unemployment produces a large and growing reserve army of labour (see Chapter 25 of volume 1 of Capital), and this reserve army keeps the real wage in check and keeps wages down.
M. C. Howard and John King have a good discussion of Marx’s wage theory as follows:
“The value of labour power [sc. for Marx], like the value of any other commodity, is given by the quantity of labour required, under the prevailing technical and social conditions, for its reproduction. The labour time needed to produce and reproduce human labour power is simply that required to keep the worker, and where relevant also the worker’s family, alive and capable of performing labour. In this sense Marx, like Ricardo had a subsistence theory of wages. He was, however, even more careful to qualify it. The value of labour power has both a natural and a ‘historical and moral element’ (Capital 1:171), and depends in part on social norms. But, like Ricardo, Marx treated this element of the wage as given in the short run, so that the concept of a fixed subsistence wage retains its validity within any given historical period.

Marx broke with Ricardo on the question of the mechanism by which real wages are maintained, in long-run equilibrium, at their subsistence level. …. Furthermore, as we have seen, Malthus’s theory of population was rejected by Marx. Instead he concentrated upon a factor specific to the process of capitalist accumulation: the industrial reserve army of the unemployed, which creates competition among workers and prevents wages from rising above the value of their labour power. This is a form of ‘overpopulation’, but it has little in common with the classical use of that concept.” (Howard and King 1985: 93).

“The existence of the industrial reserve army of the unemployed means that competition between workers for jobs prevents real wages from rising, in the long run, above the subsistence level which reflects the value of labour power.” (Howard and King 1985: 94).
So, for Marx, this rules out a long-run real wage rising and rising above subsistence level in capitalism, and that is also absolutely in line with this theory of surplus value, as I have argued here.

Howard, Michael Charles and John Edward King. 1985. The Political Economy of Marx (2nd edn.). Longman, London and New York.

Marx, Karl. 1906. Capital. A Critique of Political Economy (vol. 1; rev. trans. by Ernest Untermann from 4th German edn.). The Modern Library, New York.