Tuesday, January 28, 2014

Nouny nouns

Most of us come up with ideas which we think are good but which we don't develop or exploit. Ideas for making money or doing good, or – as in the case I am about to describe – ideas which have absolutely no possible commercial or practical applications.

Typically, we discuss these bright ideas with trusted friends or family members and get discouraged when our interlocutors are less than overwhelmed.

So let me recycle here (to the extent that I can reconstruct it from memory) one such idea which was effectively discouraged by an old academic friend and colleague whose views on the matter I may have taken a shade too seriously. Or not, as the case may be.

It relates to the topic of animism, which I raised in my previous post on this site.

There I talked about the so-called 'mind projection fallacy' discussed by Edwin Thompson Jaynes. He talked about evidence in ancient literature and pointed out that the fallacy in question would have long pre-dated written records.

We have anthropological evidence for something like Jaynes's mind projection fallacy from studies of various non-literate cultures, but my idea was to look for evidence in the structure of language.

For our natural tendency to project human-like intelligence into non-living and non-human nature is obviously reflected in various ways in the grammar and morphology of the languages we speak or know about, and these languages (would have) not only reflect(ed) but also facilitate(d) animistic modes of thinking.

You find traces of animism even in modern English idioms such as 'the wind blows', but grammatical analysis of both verbal and nominal forms takes us much further back in time.

My intention was to focus on nouns. Willard Van Orman Quine speculated (in his Word and Object as I recall) that the most basic form of noun was the mass noun – like 'sand' – rather than the count noun – like 'hill'. The former doesn't need an article ('the' or 'a'), the latter does.

But, counter to Quine's speculations, it can in fact be demonstrated by looking at the potential for inflection – grammatical suffixes and so on – of various kinds of noun in a range of languages within the Indo-European family that the prototypical noun – the 'nounier' noun if you like – is the count noun rather than the mass noun; and, of the count nouns, animate nouns are nounier than inanimate nouns; and nouns relating to humans or human-like agents are the nouniest of all.

My intention, then, was to elaborate and refine and draw out the implications of this fact: that for many languages – including some of the oldest linguistic forms of which we have any knowledge – the nouniest nouns are personal agents.

Perhaps this idea had already been developed by others at the time I first thought of it. Perhaps it has been discussed and developed more recently. Perhaps it is just not an interesting enough idea to bother with. Or perhaps none of the above applies.

Wishing, then, to maintain – at least for a little while – a state of blissful ignorance on the matter, I am deliberately postponing any scholarly delving.

I have also refrained from mentioning the name of the linguist (now in his eighties) whose work was my jumping-off point. If his name comes up in my (or anyone else's) searching it will suggest that the territory is still relatively virgin.

Sunday, January 12, 2014

Randomness in nature

I have talked before about randomness. Somehow it seems important to know whether the world we live in is driven in part by fundamentally random processes.

Some recent findings seem to confirm (though 'confirm' is probably too strong a word) what quantum theory has suggested all along: that there are basic physical processes which are truly random.

I might also mention in this context that, in doing a bit of reading on probability and related matters, I happened to come across some references to, and a paper by, the physicist Edwin Thompson Jaynes (1922-1998). Jaynes promoted the view that probability theory is an extension of logic.

This is intuitively plausible. The concept of truth (and truth tables) lies at the heart of propositional logic, and T is, of course, equivalent to a probability of 1, F to a probability of 0. Probability theory just fills in the bits in between in a quantitative way!*

Of particular interest to me is Jaynes's notion of a 'mind projection fallacy' which he sees as a root cause of much false thinking, including what he sees as the mistaken ascription of randomness to (certain) natural events or processes.

But his case seems to suffer from an overdependence on personal intuition as well as from a lack of historical perspective. For example, he develops** his concept of a mind projection fallacy without (to my knowledge) relating it to other clearly similar or related concepts – from animism to teleological reasoning – which have been widely discussed over the last century-and-a-half.

Jaynes argues that this fallacy is evident not only in the thinking of primitive cultures and amongst uneducated people but also in scientific contexts. He uses his mind projection idea to argue against certain interpretations of probability theory and statistics as well as against certain interpretations of quantum mechanics.

The basic thought seems to be that theoreticians are all too inclined to project their perspectives (their particular states of knowledge or ignorance) on to reality. He rejects, for example, the ascription by probability theorists – and physicists, it seems – of 'randomness' or 'stochastic processes' to nature. He rejects the Copenhagen interpretation of quantum theory as a mere projection of our ignorance.

But, as I say, I find it a bit off-putting that (in the cited paper, at any rate) he not only fails to acknowledge that others have developed and discussed notions very similar to his own, but also – ironically – that he seems to sensationalize and exaggerate the significance of his own insights and intuitions.

More on the substance of his claims later, perhaps.

Let me take this opportunity to thank past readers for their interest and commenters for their comments and to wish everyone a pleasant 2014.

* Like other objective Bayesians, Jaynes sees probabability theory as a formal, axiomatic system, and the calculus of propositions as a special case of the calculus of probabilities.

** Here, for example (PDF).