Sunday, September 15, 2019

Golden Oldies in Syntax

Why read the Golden Oldies?

I define a golden oldie as a paper in the framework of generative syntax written in the 60s and 70s (roughly between and including Chomsky (1957) “Syntactic Structures” and Chomsky (1981) “Lectures on Government and Binding”).

In this day and age, why should one spend time going back to these classics and reading them? Isn’t it a more economical use of time to focus on cutting edge literature: the latest Linguistic Inquiry or Syntax article, the latest syntax thesis or the latest most cited article on Lingbuzz, ignoring the detours, complexities and misconceptions of the past?

In my opinion, the golden oldies are more important than ever. Here is why:

First, the golden oldies often had interesting ideas that are still relevant today. For example, Ross' "Guess Who" is the basis for most work on sluicing done today. Also, Postal’s (1974) "On Raising" argues at length for a view of ECM constructions involving raising-to-object, which is now widely adopted in generative syntax. Even if the ideas in the golden oldies fall out of favor, they are powerful and interesting ideas, and they may be right or partially right.

To give a very short list of the kind of work I am thinking about, the following are all golden oldies: Klima (1964) “Negation in English”, Rosenbaum (1967) “The Grammar of English Predicate Complement Constructions”, Kayne (1969) “The Transformational Cycle in French Syntax (see also Kayne 1975), Baker (1974) “Notes on the Description of English Questions: the Role of an Abstract Question Morpheme”, Bresnan (1973) “The Syntax of the Comparative Clause Construction in English”, Jackendoff (1972) “Semantic Interpretation in Generative Grammar”, Van Riemsdijk (1978) “A Case Study in Syntactic Markedness” and Williams (1978) “Across-the-Board Rule Application”.

I have read all of these works, and the other cited in this post. But crucially, I have only read some of them quite recently. For example, I only read Rosenbaum (1967) and Kayne (1969) last year. Mastering the fundamentals by reading the golden oldies is a life long undertaking!

Second, the golden oldies help us to understand why the field is the way it is today, and why we do things in certain ways.

For example, the progression of papers starting with Chomsky (1964) “The Logical Basis of Linguistic theory” and Ross (1967) “Constraints on Variables in Syntax” and continuing into the 70s with papers like Chomsky (1977) "On Wh-Movement" help us to understand the importance of locality constraints (and other constraints) in simplifying transformational rules and in creating modular theories of syntax (with interacting systems of simple principles producing complex results).

Similarly, papers from that period can help understand the evolution of the division of labor between syntax and semantics, and why people take the positions they do today.

Third, and maybe most importantly, the golden oldies give one a taste of what syntactic argumentation is, and what it is to have syntactic intuitions about a problem. What does it mean to think syntactically? The argumentation in Ross (1967) (published as Ross 1986 “Infinite Syntax”) is dazzling and should be absorbed by the beginner and reviewed by the more experienced syntactician. Perhaps the particular assumptions he makes would not be accepted in a paper submitted to a journal today, but the way that he manipulates data and explores hypotheses are as relevant today as ever.

Don't more recent papers also "give one a taste of what syntactic argumentation is"? So isn't the third reason listed not so compelling? I will try to answer this in another post.

I encourage students to make sure to read golden oldies when doing research on a topic. I also encourage faculty to put such papers on syllabi and reading group lists, and to make sure your advisees have read them. These papers are an important source of insight, both methodological insight (How is syntax done?) and theoretical insight (What are the principles of syntax?).

I have been surprised that even advanced syntax students (at job interviews) sometimes have not read a particular golden oldie immediately relevant to their work (e.g., Rosenbaum's book on clausal complements). I also remember incidents where I have tried to get an advisee to read a classical paper, and was met with deaf ears such that I would have to repeat the request over several meetings. And I cannot remember the last time that I have met a graduate student who has actually read "On Wh-Movement" (or even "Syntactic Structures") on their own. So there seems to be some kind of resistance to reading the golden oldies, some idea that the work to get through the paper is not worth it in terms of what is useful in writing a modern syntax paper. 

What are your favorite golden oldies and why?

1 comment:

  1. From Pauline Jacobson (posted on Facebook 9/15/2019):

    I'm glad to hear that Raising to Object is now generally accepted (why is it still called ECM?) but your discussion of Postal 1974 implicates (perhaps you did not mean this implicature) that he first proposed it. But of course it was in Rosenbaum, 1967 under the 'it-replacement' rubric. (I am quite sure you know all this; just wanted to correct what I took as an implicature, perhaps unintended.) Postal was just arguing for that view. In fact - and here I am not entirely certain I know the history - but I think the history is that this was the received wisdom until Chomsky "Conditions on Transformations" - or maybe something a year or two earlier - where suddenly Raising to Object was discarded in favor of the ECM analysis. So it might be more accurate to say that Raising to Object is now 'reaccepted'. Then there's the question of why anyone would ever doubt Raising to Object. There have been - over the years - a few plausible empirical arguments for that (persuade vs. expect with respect to parasitic gaps, maybe?) but at the time I assume - perhaps incorrectly - Chomsky discarded that analysis because it was string vacuous, and in turn did not want string vacuous rules because the formalism for writing transformations at that moment did not allow them. I could be totally wrong on all this (someone will be sure to correct me if so). Of course as soon as movement is stated generally along with structure preservation, Raising to Object should be okay - so then additional stipulations needed to be imported to block it (e.g., all subcategorized positions must be assigned a theta role...) . Tell me if I'm wrong on any of this. In any case,my main point is that R to Obj was argued for but not proposed by Postal, and was the received wisdom until something like 1971 or 1972 or 1973 when Chomsky instead proposed the ECM analysis.


Note: Only a member of this blog may post a comment.