Thursday, May 23, 2019

Imprinting

There is an effect when students start to learn syntax that I will call imprinting: Whatever proposal is presented first to a student becomes the standard. Subsequent proposals face a burden of proof not faced by the original, in the sense that subsequent proposals need to show how they are superior to the original proposal (and why the original proposal is wrong). All things being equal, the original wins.

I call this imprinting because it seems similar to the imprinting I have heard about (but have not studied) in birds. From Wikipedia: "Lorenz studied instinctive behavior in animals, especially in greylag geese and jackdaws. Working with geese, he investigated the principle of imprinting, the process by which some nidifugous birds (i.e. birds that leave their nest early) bond instinctively with the first moving object that they see within the first hours of hatching."
(https://en.m.wikipedia.org/wiki/Konrad_Lorenz)

Here is an example from syntax: When I teach introduction to syntax (Grammatical Analysis), I try to avoid teaching phrase structure rules. Rather, I first teach constituent structure tests (e.g., the movement test, the anaphora test, the deletion test, etc.), Then once we have established the need for constituent structure, I show how it is the result of the Merge operation: Merge(X,Y) = Z. Only later, do I discuss phrase structure rules in historical context (but I do not insist students learn how to use phrase structure rules).

To make the discussion clearer, consider the structure of a transitive verb phrase. On the phrase structure based approach, one needs the rule in (1a). On the Merge based approach, one needs the operation in (1b):

1.
a. VP --> V NP
b. Merge(V, NP) = [V NP]

From experience, I believe that students who are first taught about constituent structure in terms of phrase structure rules have a hard time accepting Merge as the way to generate constituent structure. I need to explain why Merge is superior to phrase structure rules, and if the reasons I give are not found compelling, they simply stick with phrase structure rules.

Furthermore, the reasons to adopt Merge over phrase structure rules are not easy to communicate to an introductory class, where people have only had a few weeks of contact with formal/generative syntax. One of the main advantages of the Merge based approach is there is no need to distinguish between movement transformations and phrase structure rules. There is only Merge, which has two logical cases: internal Merge (movement) and external Merge (phrase structure rules). This is a beautiful and deep result, but it is not easy to convey to a class who is just starting out in syntax.

Fortunately, introducing constituent structure using Merge seems to work well pedagogically. I describe Merge using Lego blocks, which I bring to class. With Lego blocks, two Legos snap together to form yet another object (the combination of the original two Legos). Similarly, on the Merge based model, a verb "see" snaps together with a noun phrase "John" to give the verb phrase "see John". There is no notion at all in this explanation of a VP phrase structure rule.








No comments:

Post a Comment

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