Stephen van Vlack

Sookmyung Women`s University

Graduate School of TESOL

Introduction to Linguistics

Fall 2006

Week 12 - Answers for Discussion

Radford et al (1999) pp. 338-377

1. What are the different parameters that are discussed in Section 22?

-weak C -versus - strong C = C Parameter

-weak I - versus- strong I = I Parameter

-The Null Subject Parameter

-The Head Parameter

     We have mentioned that universal grammar is basically composed of two different components; principals and parameters. Principals are highly abstract rules and which apply to all languages equally and in exactly the same form. One example of this would be the principal of structural dependency. Merge would be another principal which we talked about in this class. Parameters are little bit different. All languages share the same parameters but they do not have the same values for those parameters. In generative linguistic theory it is claimed that languages will have a plus or minus value for a particular parameter and that these parameters are set quickly and once set don`t change. Parameters are further different from principals in that they're much more specific and apply to very specific pieces of structure. Once more, it has been claimed that parameters only apply to functional categories and not lexical categories (Wexler & Manzini, 1987).

     The purpose of principals and parameters, or UG in general, is to preserve and define the structure of language both as a universal and also as an individual entity. Thus, UG keeps all languages similar yet also allows for a highly constrained form of variation within functional categories. Lexical categories such as VP, NP, etc. are handled through the lexicon and not controlled directly by UG.

     If we look at the four parameters which were discussed in the book and are listed above we may concede that they certainly only would relate to functions. The categories C and I are most certainly functional categories and if we only look at the null subject parameter we find that it relates to subjects which is a grammatical function not a category. Also heads have a functional role in a phrase and not a categorial role. Linguists working in the UG tradition have come up with parameters based on observations accrued through intense and sometimes quite ingenious studies of language. Basically, I think you know this already, linguists sit down and simply dream up sentences which they think are useful or problematic for a certain theory which they have either recently devised or have been thinking about. They then ask native speaker informants to give their feelings as to whether the sentences are acceptable or unacceptable. Parameters are devised to explain certain language phenomenon.

2. How does the strong-weak parameter work to explain syntactic variation?

     OK. Now let's take a look at the C Parameter and the I Parameter more specifically. But these parameters are said to work in more or less the same way. They both revolve around the idea of strength or weakness within a specific functional category. The basic idea is that if the category is weak than it does not need to be filled but if the category is strong then it needs to be filled. The simplest way to understand this is to compare it to something like a vacuum. If CP, for example, is strong as the authors of the blue book claim it is, then it will pull something into it under the right conditions.

     English supposedly has a strong C and weak I meaning that, when necessary, C must be full but I need not be filled. Since I is claimed to be weak, it is OK for I to be left empty. These parameter values have been created and used to explain utterances like the ones directly below.

(1) What did the porcupine do?

(2) What are you doing?

     The interesting phenomenon in both of these utterances is that the main verb 'do' does not undergo any kind of movement and a so-called dummy auxiliary is inserted into the structure and winds up in C position. This becomes more interesting when we studied closely related languages like German and more particularly older versions of English itself. In looking at the older versions of English we find that this was not always the case. In early modern and Middle English main verbs were able to move directly into CP position in order to form interrogatives. Now we need some sort of explanation for why this no longer happens in English. The first thing that needs to be done is to explain why the verb doesn't move. The explanation for this, according to the book (and this is a very mainstream explanation), is that IP is weak in English and as such will not pull the verb into it. So the verb remains where it is. Now, this creates a problem because even though IP is weak CP is strong and CP requires something move into it in order to satisfy this need. In order for this to happen a dummy auxiliary is created simply to occupy CP position as dictated by the parameter. That is basically why and how that works.

     For us, we have been working with a slightly different version of the scenario and more specifically one in which IP is replaced by two different functional categories, namely AGRP (Agreement/Aspect Phrase) and TP (Tense Phrase). The version we been looking at is very different from the one described in Radford et al. (1999) because for us we have been basically talking about verbal raising as an important type of movement and which is necessary for the verb to acquire its tense and agreement features. In the Radford et al. (1999) version IP is weak and does not allow things to move into it and thus the only way for the verb to acquire its tense agreement features is for them to trickle down out of IP (which is the holder of these features) and onto the verb. Now, from a very general point of view it is probably not that big a difference whether the features trickle down or the verb moves up to get them. Either way the surface level word order is the same so the result is the same even though the process might be different. The difference does become important, however, when we begin to look at the I Parameter. According to the I Parameter verbs either do or do not move forward and in English they are not supposed to, but we have the verb moving forward in our AGRP/TP analysis. This becomes a problem because if the verb moves forward to TP and then then AGRP what is to stop it from moving forward into CP when necessary (that is, for interrogatives). In order to make our theory work we need to come up with some sort of barrier which allows verbs to move into AP and TP but would stop them from moving into CP. The solution that has been put forth is to argue that TP would be a barrier and would stop the verb for moving because the features of verb as a lexical item would not be able to carry over into the CP node. This TP as a barrier analysis has a precedent in what we already saw with exceptional clauses.

(3) The porcupine_i wanted [John_j to bite himself_j].

(4) *The porcupine_i wanted [John_j to bite itself_i].

(5) The porcupine_i wanted [PRO_i to bite itself_i].

In the above examples (3) and (4) have complement clauses that begin with CP and therefore also contain AGRP and TP both. In (5), however, the complement clause is a so-called exceptional clause which does not contain the levels of CP and AGRP, but rather starts with TP. We can see that (5) allows features to move from the first clause across to the complement clause. That is, there is obviously no barrier to these features moving. The same does not seem to apply to (3) or (4) as shown by the unacceptability of (4). this means that there must be sum sort of barrier which stops the features from carrying and we can posit AGRP as just that barrier.

     It is not imperative that you know all the details of all these different arguments. What is important is that you have a general understanding of what these parameters are used for and how linguists go about forming theories.

3. Why did OE (Old English) and EME (Early Modern English) have a null subject while present day English cannot?

     Well, the obvious answer to this question is because of the parameter settings. Linguists have been trying to come up with reasons why some languages Pro-drop and others do not. The reasons they give in the chapter of the blue book are neither clear nor valid. German, for example has a complex agreement system, like that of Old English, in which the subject is predicable from inflectional endings marking verb agreement, yet it is not a prop-drop language. Korean and Chinese on the other hand have no person or number agreement features marked on verbs (even less than English) yet they are both pro-drop languages as are many of the languages in the region here despite historical affiliation. So, it seems there are no fast and easy ways to determine why some languages drop `pro` and other do not.

4. What is Logical Form and what is it used for?

     LF is composed of interpretations of words combined with interpretations of grammatical structures (functions). The purpose of LF is to match the syntactic form with the intended meaning and to create structures that are not ambiguous. It is extremely important to remember that at the deep structure level when utterances are first put together they are tremendously ambiguous both at the syntactic and the word level. (Remember when we were going over lexical items we said that lexical items are ambiguous, that is they don't have any particular meaning, until they are inserted into a particular structure and context.)

     LF is the point where at syntax and semantics meet. It is like the elbow joint which joins them and makes them both work in relation to each other. Again the idea being them only go to form utterances we first formed into structures which are highly ambiguous. From the deep structure anyone of hundreds of different surface structures can be formed. It is logical form (LF) which determines which of these surface structures will be created. LF does this by matching intent with structure. It actually quite simple. LF takes the basic structure and fine-tunes it in order to reduce ambiguity and create a structure which does want the speaker wants to do. That is LF uses structure as well as lexical forms in order to match the surface structure with the intent of the speaker. Simple!

5. How does Logical Form relate to covert movement?

     When we think about covert movement we just need to think that there is movement and there is other movement. Covert movement is the other type of movement. Covert movement is movement which occurs either sight unseen (that is, the movement occurs but it doesn't actually show in the surface structure) or the movement occurs at the level of LF (that is, LF forces a certain movement to occur which actually violate some parameters). It is never posited that movement at the level of LF must occur later than movement which occurs for structural (parameter enforced and regulated) purposes. That is the basic idea. Again, is important to realize why linguists do this but you do not need to know exactly how all this covert movement works.

Section 22 - Syntactic Variation

The main thing that the authors are trying to show us in this section is that we have to go a little deeper into our descriptions to look at how individual languages and even dialects within a language superordinate vary. Basically, they are trying to account for differences we see in the behavior of languages in as simple and accurate a manner as possible. In order to do this, they mention three different parameters that further apply in addition to the syntax we have already learned.

Two of these parameters are basically the same. They revolve around a strong/weak dichotomy in certain projections on our trees; namely, C and I. The basic difference between a strong and weak category is that a strong category must be filled if possible and has the strength to actually pull things into it to fill it. Weak positions need not be filled and have no means at their disposal for filling themselves.

This distinction can easily explain differences between Standard American Vernacular English (SAVE) and the Early Modern English (EME) of Shakespeare. In SAVE, I is weak, while C is strong. This accounts for the fact that questions in English are formed by moving something into C, but things cannot be moved into I position unless they have to be. Main verbs cannot function as questions in SAVE because they are blocked from moving into I since I is weak.

1. *Walked you to the station?

Is not an acceptable sentence in SAVE, but it was so in EME because here the I was strong. It pulled the main verb into it on the slightest pretext and from there it was further moved into strong C. This still doesn`t explain why main verbs can`t function as questions in SAVE. Why can`t we just move them from V to C? Well, there is a movement constraint called `head movement constraint` which says that we can only move a head into the next phrase higher than it. The simple problem in SAVE is that movement from V to I is blocked because I is weak and we can`t just move things from V to C because they are too far away from each other.

The other parameter that was discussed in the section was null subject. We stated much earlier that all clauses have to have subjects, but in reality not all do. Many languages drop their subjects. The basic idea is that we can drop the subject if it is somehow marked elsewhere, like the verb morphology. Languages which have heavily inflected verbs, verbs which clearly show the person and number of the subject, need not overtly mark the subject. The subject is rendered null because it is simply not necessary to say the same thing twice

Section 23- Logical Form

Logical form is the link between syntax and semantics. It figures in as one of the levels of representation in between the UR and the SR of utterances. Logical form is the meaning of sentences that is related to a kind of syntactic representation (see example 392 on page 362 of the text).

The reason that linguists think we need Logical Form (LF) is that there is often ambiguity in language. It is the job of LF to map form to meaning. What pragmatics tells us is that any utterance is potentially ambiguous. What we need is some indicator to map out to us what the intended meaning of the utterance is so that we can map the intended meaning to the syntactic form. This is done through LF. The example they give us in the text is a good one, so I will use it here.

2. Frank spotted the man with the telescope.

The two possible readings of this utterance are; 1. Frank used a telescope to see the man, as he was very far away, and 2. the man Frank saw had a telescope.

When we go to make this utterance, we are somehow aware of the ambiguity and let logical form help us map out (or direct) the syntactic structure that we need to create the correct utterance.

Following all this LF stuff, syntacticians further proposed that there must be some sort of covert movement which allows us to better match meaning and form. For the most part, covert movement was created to explain how quantifiers, like every, work in English. Since we assume that quantifiers work basically like wh-words and are moved, according to operator movement, into the specifier position of C. This, however, creates problems for a system of syntax without convert movement.

Due to the crossover principle, we can not move an operator across a co-referenced (co-indexed) pronoun. Yet this happens.

3. His hamster loves every boy.....

4. Every boy his hamster loves......

Both of these are acceptable and it is plain to see that in the second one the operator `every hamster` has been moved to C specifier position. This is in violation of the crossover principle. We, therefore, must assume that the movement was covert. We also assume that the movement was created due to the influence of LF on the structure of the utterance.

Lee (2001) Chapter 11, Cognitive Linguistics and Discourse Analysis

6. How do frames relate to discourse?

     The basic idea underlying frames and discourse is that all people use these frames as a way of structuring certain elements of their discourse. The linguistic structures people put together are determined by what they know and how they think. Such knowledge is seen in the cognitive approach as being organized in frames. Now, for communication to ensue it is imperative that people are using the same or similar frames otherwise they will not be able to understand each other. It is however also true that frames are build on the brunt of the experience one has had in the world either directly or indirectly. One of the beauties of language is that it enables us to constantly redesign our existing frames and even build new ones. Thus, no two people have exactly the same frames. It should be noted that frames are immensely complex entities which interconnect. Therefore, it is to be expected that in discussing similar concepts people will do so while containing different elements in their frames. This can obviously lead to miscommunications and misunderstandings. People can misunderstand each other because they are operating with different elements in their frames. This is a rather general problem and usually an innocent one.

     One argument frequent used by generativists against the cognitive approach is that it seems unconstrained particularly in the way it addresses the idea of construal and word meaning. If people have linguistic systems based entirely on experience, so the generative argument goes, and experience varies then we would expect much more faulty communication to occur than has been observed by researchers. Generally speakers of the same language are seen as understanding each other without great difficulty or conflict. The generativists have interpreted this fact as being a clear indicator that there is some sort of overriding genetic system determining how we communicate. UG takes our somewhat diverse experience with the world and makes it linguistically the same for all native speakers of a language.

     Cognitive linguists stand by the validity of their model citing several major arguments supporting the belief that people do in fact approach situations from different points of view and with different linguistic structure, but still somehow manage to communicate effectively. The first of these arguments is that there do in fact seem to be more misunderstandings in the regular process of communication than is claimed by the generativists, but they are subtle. Thus, subtle miscommunication is a major part of all linguistic exchanges. At the same time cognitivists also claim that there are in fact mechanisms in place which serve to reduce miscommunication. The first of these is the widely recognized phenomenon of speech accommodation. Speech accommodation basically claims that as people speak together their speech will become more similar sometimes in a marked way. This is a subconscious process and seems to apply universally as a result of brain structure and more specific ally frequency and recency effects. Thus people are able to understand each other more because their speech patterns become more similar with contact. The more contact the more similar their speech patterns. This similarity in speech patterns has a direct effect on both frames and radial categories because the major way in which these are built is through linguistic experience. For humans, it should constantly be borne in mind, the main way in which indirect learning takes place is through language. Thus, language is our major tool for learning in general. The last argument of the cognitivists is that language is also made similar, and communication problems are thus avoided, by the simple fact that language is composed of many conventions. These conventions are learned and become coded in the brain as chunks. Many linguistic phenomenon exist simply because they are conventions, these are such things as collocations and constructions. Conventions make up the brunt of much linguistic behavior and since most speakers share these conventions their language is much more similar than diverse. Use of language revolves around the constant use of these conventions with a certain amount of extension.

7. How do radial categories relate to discourse?

     How radial categories are used in discourse is a bit more complicated because there may be a degree of intent in the miscommunication as a result of foregrounding or focusing perspective on certain aspects of a construal in an ongoing conversation. This is made possible by the believe in cognitive linguistics that radial categories contain a multitude of various, but related, meanings of the same word unit. The central idea to remember in the cognitive model is that just because people are using the same word unit does not mean that they are using the same exact of the word. Their intent might be different as made possible by the concept of radial categories. When people speak they choose carefully among multitude of possible meanings which one might be best for their current situation. There is some degree of intent in choosing meaning and the specific meaning chosen might vary from the other person or persons involved in the conversation. Misunderstandings can also take place not due to specific intent but innocently if the interlocutors do not have shared meanings in their radial categories or if they divide up similar concepts differently, as is often the case in cross-linguistic contact.