[ print ]

Professor of Linguistics and Western Civilization, Columbia University; Cultural Commentator; Author, Doing Our Own Thing
Are subordinate clauses more typical of languages with a long literary tradition than integral features of human speech?

Contemporary linguists tend to assume in their work that subordinate clauses, such as "The boy that I saw yesterday" or "I knew what happened when she came down the steps", are an integral part of the innate linguistic endowment, and/or central features of "human speech" writ large. Most laymen would assume the same thing. However, the fact is that when we analyze a great many strictly spoken languages with no written tradition, subordinate clauses are rare to nonexistent. In many Native American languages, for example, the only way to express something like the men who were members is a clause which parses approximately as "The 'membering' men"; the facts are similar in thousands of other languages largely used orally.

In fact, even in earlier documents in today's "tall building" literary languages, one generally finds a preference for stringing simple main clauses together —she came down the steps, and I knew what happened rather than embedding them in one another along the lines of when she came down the steps, I knew what happened. The guilty sense we often have when reading English of the first half of the last millennium that the writing is stylistically somewhat "clunky" is due largely to the marginality of the subordinate clause: here is Thomas Malory in the late fifteenth century:

And thenne they putte on their helmes and departed 
and recommaunded them all wholly unto the Quene
and there was wepynge and grete sorowe
Thenne the Quene departed in to her chamber
and helde her
that no man shold perceyue here grete sorowes

Early Russian parses similarly, and crucially, so do the Hebrew Bible and the Greek of Homer.

At the time that these documents were written, writing conventions had yet to develop, and thus written language hewed closer to the way language is actually spoken on the ground. Over time, subordinate clauses, a sometime thing in speech, were developed as central features in written speech, their economy being aesthetically pleasing, and more easily manipulated via the conscious activity of writing than the spontaneous "on-line" activity of speaking. Educated people, exposed richly to written speech via education, tended to incorporate the subordinate clause mania into their spoken varieties. Hence today we think of subordinate clauses as "English", as the French do "French", and so on — even though if we listen to a tape recording of ourselves speaking casually, even we tend to embrace main clauses strung together in favor of the layered sentential constructions of Cicero.

But the "natural" state of language persists in the many which have had no written tradition. In the 1800s, various linguists casually speculated as to whether subordinate clauses were largely artifactual rather than integral to human language, with one (Karl Brugmann) even going as far as to assert that originally, humans spoke only with main clauses.

Today, however, linguistics operates under the sway of our enlightened valuation of "undeveloped" cultures, which has, healthily, included an acknowledgment of the fact that the languages of "primitive" peoples are as richly complex as written Western languages. (In fact, the more National Geographic the culture, the more fearsomely complex the language tends to be overall.) However, this sense has discouraged most linguists from treading into the realm of noting that one aspect of "complexity", subordinate clauses, is in fact not central to expression in unwritten languages and is most copiously represented in languages with a long written tradition. In general, the idea that First World written languages might exhibit certain complexities atypical of languages spoken by preliterate cultures has largely been tacitly taboo for decades in linguistics, generally only treated in passing in obscure venues.

The problem is that this could be argued to bode ill for investigations of the precise nature of Universal Grammar, which will certainly require a rigorous separation of the cultural and contingent from the encoded.

JOHN H. MCWHORTER is Assistant Professor of Linguistics at the University of California at Berkeley. He taught at Cornell University before entering his current position at Berkeley. He specializes in pidgin and creole languages, particularly of the Caribbean, and is the author of Toward a New Model of Creole Genesisand The Word on the Street : Fact and Fable About American English. He also teaches black musical theater history at Berkeley and is currently writing a musical biography of Adam Clayton Powell, Jr.