This theory is certainly very ingenious; it is only unfortunate that it does not accord with facts, as far as our observations can trace them. You have shown that the comprehensive compounds of the Delaware idiom are formed out of other words expressive of single ideas; these simple words, therefore, must have been invented before they were compounded into others, and thus analysis presided over the first formation of the language. So far, at least, Dr. Smith’s theory falls to the ground; nor does he appear to be better supported in his supposition of the pre-existence of generic terms. For Dr. Wistar has told me, and quotes your authority for it, that such are seldom in use among the Indians, and that when a stranger pointing to an object asks how it is called, he will not be told a tree, a river, a mountain, but an ash, an oak, a beech; the Delaware, the Mississippi, the Allegheny. If this fact is correctly stated, it is clear that among those original people every tree is not the tree, and every mountain the mountain, but that, on the contrary, everything is in preference distinguished by its specific name. Peter S. Du Ponceau to John Heckewelder, 21st August, 1816, Letter XX, in History, Manners, and Customs of The Indian Nations who once Inhabited Pennsylvania and the Neighbouring States. 406-407 (1881)
I was alerted to the the quoted passage by p. 137 in James Turner's entertaining Philology: The Forgotten Origins of the Modern Humanities, which is (recall) prompting a number of these digressions. I was elated to read Turner and Du Ponceau. Let me explain. Smith's "elegant treatise" is now fairly obscure. For reasons I will never understand, nor approve, the editors of the most widely used modern edition of Adam Smith's works, the so-called Glasgow Edition of the Works of Adam Smith, moved Smith's "very ingenious" essay on the origin languages (hereafter Languages) from its original role as an appendix to The Theory of Moral Sentiments (hereafter TMS)-- where it appeared from the third (1767) edition onward -- to an appendix of the student notes to Smith's Lectures on Rhetoric and Belles Lettres (largely ignored by contemporary philosophers). This means that modern readers of Smith's moral philosophy get a truncated version of his intended moral psychology and his more general system.
I learned from Glory Liu's wonderful forthcoming work on the American reception of Smith that there were serious readers of Smith's TMS. So most early American readers of Smith could have known of Languages through it.
In context, Turner treats Adam Smith as a representative of the Enlightenment tendency toward baseless conjectural speculation that retarded the development of American linguistic and anthropological science (see also p. 144). Du Ponceau remark is interpreted with quotes from another American philologist John Pickering as an important signal of the "sober," more scientific, inductive mindset focused on the collection of "data" about languages in comparative fashion. Now, I note this because some other time I want to turn to the peculiar and to my mind regrettable feature of Turner's larger argument which celebrates the slow discovery of historicism as the great achievement of scientific philology, while treating the rival conceptions of philology and science in whiggish terms. But that's for another time. Unfortunately, Turner is not really interested Du Ponceau's own argument.
Du Ponceau's argument relies on a tacit premise: that the languages spoken by American 'savages' are, in a conceptual and temporal sense, closer to the original language(s) spoken by the first inventors of language. So, to study then contemporary 'savages' is then to be offered a picture into the immaturity of mankind. The premise is, as Chris Berry has shown, indebted to Lockean anthropology widely used in eighteenth century 'science of man' and familiar to modern scholars through Ferguson's works. This premise, of course, fed into a more noxious colonial ideologies.
But while there are Lockean strains in Smith I argue in my book against Berry (44-45) that Smith rejects this premise. In Languages, Smith clearly has "early" savages in mind not then-contemporary ones. Since Du Ponceau alerts the reader that he is quoting from memory, the natural reading of his letter is that he conflated the striking example from Smith as purported evidence for a more familiar theory that relies on the Lockean premise (mistakenly attributed to Smith). There are two further bits of evidence for my reconstruction.
First, that Du Ponceau is misremembering Smith is clear from an important fact unremarked upon by Turner. Du Ponceau explicitly takes himself to be refuting with empirical data an argument for the synthetic theory of language (to be opposed to an analytic theory that he endorses). But while Smith uses such a distinction in other places, it is not part of his theory of language let alone its origin. And, in fact, in context Du Ponceau explicitly notes that applying the analytic-synthetic distinction to the origin of language is his own (""as I call them" (p. 405)) not Smith's.
Second, the so-called analytic theory is couched in terms familiar from Locke's philosophy: in it comprehensive compounds...are formed out of other words expressive of single ideas. This seems the doctrine of Locke's chapter (XII) on Complex Ideas (see also Ch. XIV). Smith, nor Hume, never use 'single ideas" in this sense, preferring 'simple' ideas.
As it happens, despite the reputation for offering conjectural history, Smith's theory of the origin of language does rely on empirical data. It's data is not derived primarily from comparative linguistics,+ but rather from the (quite perceptive) observations of children's babbling.* Smith is relying on another kind of premise, one that later came to be associated with (the now discredited) Haeckel's Biogenetic law, that early childhood development recapitulates the development of the species.
So, one natural way to interpret what's going on here is two kinds of data being brought to bear with the help of auxiliary premises derived from a much broader anthropology on a topic (the origin of language) that may not seem especially amenable to empirical enquiry.+ It's not good vs bad science. It's two kinds of science, both relying on limited data.
I could stop here, and use this as a morality tale about the way Turner is in the grip of a certain kind of scientific success story (philology's discovery of Indo-European language family) and so mis-represents the underlying history. (This is part of my larger gripe about his book.)
But life turns out to be more complicated in two distinct ways. First, recall that Smith interprets the data in order to secure another claim, that the development of language and mind requires the development of certain kind of mental operations; each such operation he calls an abstraction (from the senses), and these operations of abstractions are themselves measure for the mind's capacity for metaphysics. And the child's slow development of abstraction is, then, a proxy to study the slow development of metaphysics in the species and, thereby, the identification of a kind of natural metaphysics common to adult members of the species (uncorrupted by learning). [And, in a further twist, this natural metaphysics is not true.] I have told this detailed story in my book.
And, second, it turns out De Ponceau introduces the distinction in order to discuss a different debate altogether, namely which kind of language (an analytic or synthetic one) "is the most perfect or is preferable to the other." (405; recall also my treatment of classics last week; Turner misses this.) De Ponceau goes on to write:
In the course of my reading, I have often seen the question discussed which of the two classes of languages, the analytical or the synthetical (as I call them), is the most perfect or is preferable to the other. Formerly there seemed to be but one sentiment on the subject, for who cannot perceive the superiority of the Latin and Greek, over the modern mixed dialects which at present prevail in Europe? But we live in the age of paradoxes, and there is no opinion, however extraordinary, that does not find supporters. To me it would appear that the perfection of language consists in being able to express much in a few words; to raise at once in the mind by a few magic sounds, whole masses of thoughts which strike by a kind of instantaneous intuition. Such in its effects must be the medium by which immortal spirits communicate with each other; such, I should think, were I disposed to indulge in fanciful theories, must have been the language first taught to mankind by the great author of all perfection.
All this would probably be admitted if the Latin and Greek were only in question: for their supremacy seems to stand on an ancient legitimate title not easy to be shaken, and there is still a strong prepossession in the minds of the learned in favour of the languages in which Homer and Virgil sang. But since it has been discovered that the barbarous dialects of savage nations are formed on the same principle with the classical idioms, and that the application of this principle is even carried in them to a still greater extent, it has been found easier to ascribe the beautiful organisation of these languages to stupidity and barbarism, than to acknowledge our ignorance of the manner in which it has been produced. Philosophers have therefore set themselves to work in order to prove that those admirable combinations of ideas in the form of words, which in the ancient languages of Europe used to be considered as some of the greatest efforts of the human mind, proceed in the savage idioms from the absence or weakness of mental powers in those who originally framed them.--Letter 20, (pp. 405-6)
And because Du Ponceau assumes that Smith is working with Locke's anthropology, he treats Adam Smith as an exemplar of the debunking kind of philosopher who shows that what was once taken as evidence of civilizational ingenuity can be explained by the operation of rather minimal mental powers. To put this in terms familiar of Dennett, Smith is treated as somebody who uses cranes to explain the appearance of skyhooks.
As it happens, Smith is committed to the idea that over time languages develop from a kind of original simplicity of expression to a more complex form of expression. And he is committed to explaining this by way of the development of mental operations. (He compares this to the evolution of a machine.) But he also thinks that over time as nations and languages intermingle (which he approves of on political grounds), the "modern mixed dialects," languages also get simplified (in the manner he takes to occur among pidgins): "And thus upon the intermixture of different nations with one another, the conjugations, by means of different auxiliary verbs, were made to approach towards the simplicity and uniformity of the declensions."** So, Smith simply would reject the way Du Ponceau sets up the problem even though Du Ponceau recalls correctly some of the most striking features of Smith's theory.
What does this have to do with early anthropology? Not much except that Du Ponceau's letters are published as an appendix to Heckewelder's book (1819), which is among the first works in which we can discern the methodology we now associate with scientific ethnography. But about that some other time.
*I call it perceptive because Smith recognizes that there is a critical period for language acquisition.
+As it happens Smith is mis-represented in another way. For, Smith also relies on the comparative method characteristic of philology--he appeals to comparative claims about Latin, Greek, and Hebrew, too.
Comments