Fossilized Syntax

A whole range of expressions, known technically as root small clauses, may be living fossils, vestiges of a period before speech used full sentences, Dr. Ljiljana Progovac of Wayne State University told a conference of the International Linguistic Association this past March 31.

For many American’s  Mad magazine’s What me worry? slogan is the best known American language example of a root small clause (RootSC). If her hypothesis is true, it will go a long way toward justifying the study of speech origins and its evolution. An evolutionary account, Dr. Progovac says can “shed light on the very nature of syntax, including why every sentence is built upon a small clause … [and] why [root small clauses exist today] only in marginalized roles.”

What me worry? illustrates several common features of root small clauses.

• The verb has no tense. In a full sentence, this slogan might read “What you expect me to be/act/feel worried?” So the sentence has not just been shortened. The verb has no tense marker.
• There is no nominative case. Alfred E. Neuman means he’s not worried but he uses the accusative me rather than the nominative I. (Technically, the case is the “default” case, the one used unless a rule is followed. In English the default case is accusative; e.g., it’s me; t’was him. In many languages the default case is nominative and their RootSCs use it.)
• What doesn’t force a changed word order. If Alfred E. Neuman were less of a poet, he could just as easily have said Me worry? In English the use of a question word usually forces a movement in word order: He is worried about his mortgage but What is he worried about? The use of what forces the subject to move to a new point in the sentence. Bill worries becomes What worries Bill? But no such move is required in RootSCs. You may protest that the Mad magazine slogan is as much an exclamation as a question, but that’s the point. You cannot use RootSCs to form a question of the kind that can be asked with full syntax. The question words can be used to generate small clauses: Who, Bill?, Why me? and they can just as easily be used without the question word: Bill? Me? By not forcing a move, small clauses cannot express things available to users of full syntax.

Taken as a whole (tenseless verbs, default case of nouns, no moves to alter meaning) root small clauses appear to be expressions with only a partial syntax. Yet they do have some rules because the words cannot be ordered randomly. (No: what worry me; worry me what; me worry what; me what worry; worry what me.) And despite the limitations many things can be stated in the form of a RootSC. Progovac maintains that many of today’s syntacticians agree that every sentence is built on a small clause.

If that last claim is true, it would mean sentences are generated by a computation that first produces a core, then merges tense and case markings with the core, and finally moves the subject as necessary. For example, a core sentence might be Me in Rome. After the merge operation marks the verb for tense and gets the case right, the sentence might be I am in Rome. Finally, a move operation changes that to Why am I in Rome?

This process seems unduly complicated and logically unpersuasive; however, if viewed in evolutionary terms it makes good sense. If first we developed the ability to speak in core clauses and then evolved the ability to merge syntactical markers and finally evolved the ability to move subjects, we have a process that can emerge bit by bit over many generations. The alternative is a proposition I have often protested on this blog, that the whole syntactical marvel appeared in one swoop.

It might seem that the transformation from Me in Rome to Why am I in Rome involves more than adding syntactical sophistication. A new thought seems to be expressed, but that is not necessarily so. (Here, by the way, I’m meditating on my own, building on hints I picked up from the Progovac presentation. She may wish to disavow my speculation.) Tone of voice could put the small clause into interrogative, doleful form. We tend to read Me in Rome as an assertion, but it can be read aloud as a puzzled lament of wonder. Blog visitors can test their own acting abilities by reading the clause aloud in a way that sounds more like a question than a statement of fact. I am in Rome can certainly be read aloud in a tone of dismay and wonder. Thus, the evolution of syntax need not be understood as an evolution in the capacity to perceive and respond to perceptions. It could have been part of a shift from an emotion-based expression in which understanding depended almost entirely on tone of voice to a coherent expression in which understanding depends mostly on syntactical guidance. (Note: even today tone of voice can replace move to form a question: Am I going to die? versus I am going to die?) Progovac does point out that even today we use exaggerated intonation when using RootSCs to express incredulity: e.g., JOE [caesura] aFRAID!

As we students of Jean-Louis Dessalles know, this kind of evolution cannot be thought of in terms of some natural progress. Each stage of evolution is quickly optimized for the needs and circumstances of a species. Progovac does not speculate on the functional changes that demanded syntactical changes, and we won’t either. (At least not today.)

She does speculate on the stages of syntactical evolution:

Pre-syntax: A protolanguage or pidgin stage in which there is no syntax. Instead one image follows another. Me. Rome.

Two-Word Syntax: A combinatorial “breakthrough” in which two words express a single image: Me Rome. This shift introduces predicates. The pre-syntax can be understood as positing one figure (me) followed by another (Rome). Two-word syntax tells us something about a figure with a hint of a background as well (me is in Rome).

She continues, but I don’t think I will. The point for this blog is that she is proposing a transition stage (or stages) between protolanguage and full speech, and she is pointing to some empirical evidence to back up her claim. Many (all?) languages include RootSCs to this day. They are especially useful when expressing things like incredulity (Who me?), and wishes (Me first), elements of a kind of timeless here-and-now, to be a little oxymoronic.

In the past, if we noticed these sorts of expressions, we might have thought of them as merely pithy abbreviations, but the use of default case and tenseless verbs argues that something else is going on.

Comments

[Ken Arneson]

It's interesting to note that there are often relics of evolution in the stages of child development (e.g. human embryos have gills and tails at an early point). These three stages--a one-word stage, a two-word stage, and then a sudden leap into full adult syntax--also appear in human child development. It seems reasonable to speculate that we each develop through those particular stages in order, because we evolved through those stages in order.
--------------------------
BLOGGER: Although this is a reasonable speculation, we will need careful empirical data before embracing it. But at least now we have a good question to ask.

[TLTB]

While I don't think it is consensus that all sentences start with a small clause, most syntacticians would not consider it an outrageous assumption. There is good evidence that many sentences begin with small clauses; they are certainly accepted as one of the species in our linguistic kingdom.

But there is little here to support a gradual evolution of the properties of syntax other than the three stages you note: no syntax, two-word syntax, and full-fledged syntax. The gap between the second two still looks huge.

What Chomsky and others have suggested is that we get from stage two to three with one simple adaptation: recursion. We go from being able to merge two elements (two-word stage) to be able to merge two elements and then merge the object formed of those two elements to a third element (and so on). It is a plausible scenario that the other properties we get: syntactic case, movement, agreement, etc. could be the result of this one adaptation. Take movement, e.g., Chomsky has argued that if you have a productive operation of merging objects together, you can also merge an object that has already been merged (what he called 'internal merge'). So you could merge a wh-phrase like 'why' at the front of the clause even after it has been merged down at the base of the clause. This yields the property of movement and it comes 'for free' with the adaptation of productive merging. Case and agreement systems might then have emerged (perhaps very quickly) as a response to the presence of movement in syntax. After all, if you can move things out of where they should be, you need a way of tracking their movements. This is one of the things that case and agreement do.

Nice post.

[Victor]

You have very nice site!
http://www.volny.cz/coachshoesitaly/
http://idisk.mac.com/louisvuittonsunglass/Public/index.html
http://amping-water-heaters.blogspot.com