The essential idea in this question is that one uses the tests in order to construct a tree that makes sense according to the tests carried out. Nevertheless, many many people included nodes in the tree that they had not tested. That is, only obvious constituents like ‘in the park’ were tested, while others, such as ‘often ate mushrooms in the park’, ‘ate mushrooms in the park’ or ‘mushrooms in the park on Wednesday’, etc. very rarely appeared. Despite not testing these, most people managed to magically include a correct VP node in their tree, showing that they were not building the tree exclusively from the test results.

A good example of this was the NP that was sometimes suggested consisting of:

“mushrooms in the park on Wednesdays”

How do we know that this makes absolutely no sense as a NP? Well, one test would be by movement: particularly movement that might be helped by putting the sentence into the passive. In a passive sentence, the NP (formal) that is the direct object (functional) is moved to become the subject (functional) constituent: which, if you had this entire phrase as an NP, would require that the passive be:

Moreover, mushrooms in the park on Wednesdays were often eaten by Susan.

And if you think that is a good English sentence, it is back to Sprachpraxis with you! This is important because it shows why certain phrases work in English and others do not. It is not a matter of style or of a sentence being less good than it might be: this is simply a violation of the grammatical structure of English.

For your marks, the more evidence that you gave in your answer that you were putting a node in your tree because of concrete test results, the more marks that you got! This was even the case when you got it wrong: if you said that something followed because you had got the test results wrong, but it would have followed if the test really falls out that way, then this showed that you were using the test results. The point of the question was not then to get the most accurate tree that you could on ‘general’ principles or what you have memorised as being a ‘good’ tree, but to show that you can motivate phrase structure on the basis of the tests; this should be the reason that is always used actually! Phrase structure rules, such as S -> NP VP, are nothing more than a record of all the tests that linguists have been trying out for the past 40 years on sentences of various kinds.

When you do not understand why a particular rule should be as it is, for example, NP -> Spec N′, then it just because you are not familiar with all the sentences that linguists have been applying the tests to! And that is not a problem as long as you can apply tests yourself; these phrase structure rules are then a record of a lot of detailed work about how sentences are put together and so, indirectly, about how language works and how it does not.

This question was to let you show what you had learnt about historical linguistics. The particular examples that we used in class and in the readings usually involved the development of the modern languages (Spanish, German, English, etc.) via various consonant shifts from Proto-IndoEuropean. Positive marks were obtained by mentioning this fact, for mentioning Grimm and Grimm’s law, for listing the effects of the consonant shift, etc. Positive marks were obtained for concrete content, not for general statements, such as, e.g., ‘there was a sound shift’ or, worse, ‘the letters changed’! Since the question explicitly asks about the creation of new languages, variation within English, as some of you decided to write about, was generally not an appropriate answer. The Great Vowel Shift, etc. had nothing to do with the development of modern languages from Proto-IndoEuropean, as some of you suggested. The only exception to that was the clever suggestion that the Great Vowel Shift contributed to the existence of a separate language of Scots English, which is also true!

The three main stages of the English language were Old, Middle and Modern. Just naming these was, of course, not enough—what had to be provided were examples of the linguistic features of these stages of the language that allows us to separate them so clearly. Mentioning that Old English was a Germanic language because of being brought to England by the Angles, the Saxons, the Jutes, etc. was good, saying just what the features of that Germanic language were, (e.g., inflection for case and gender, SOV word order, etc.) was better. The same for Middle English and Modern English: introduction of the use of ‘do’ in questions, etc., dropping of inflections (which?), change in word order, etc. The usual historical events to cite are: the original invasion itself (around 450 AD note, some people put it substantially earlier!), the prolonged interaction with the Vikings, the invasion and conquest by the Normans, etc. Again, marks were given for concrete examples, not for generalities.

First, it was necessary to distinguish the four sounds completely. This was fairly straightforward: [g] : +velar, +plosive, +voice vs. [k]: +velar, +plosive, -voice; [p]: +bilabial, +plosive, -voice vs. [b]: +bilabial, +plosive, +voice. Note that using non-technical terms here, such as ‘soft’, ‘hard’, ‘with more air pressure’, etc. did not get any marks because these terms are so vague as to be useless. They were also often used in the wrong places, showing that there was not really an adequate understanding of what was going on. The difference between [p] and [b] is nothing to do with ‘blowing harder’ for example, as some people almost wrote. Additional marks were obtained here if you showed that you knew that these features were distinctive features, i.e., features that serve to distinguish the phonemes involved.

For ‘pin’ and ‘spin’ either good introspection or broader knowledge of pronunciation in English was tested: the difference is usually given as ‘pin’ using an aspirated ‘p’, written [p^h], and ‘spin’ using a non-aspirated form [p]. Aspiration means that a small burst of air follows the release of the lips: it is not a distinctive feature, but is another example of an allophone. Some of you noticed that there was a difference between pin/spin and described the latter ‘p’ as resembling a ‘b’. This was quite good as it showed that even if you did not know about aspiration, you could notice the difference between the ‘p’ in the two phonetic contexts. An unaspirated ‘p’ at the beginning of a word (there are no such sounds in English, although some other languages do this!) would often be misheard as a ‘b’.

Most of you had come across the example with the different kinds of /l/ before and so were able to answer this part of the question quite well. This also led on to some good definitions of allophones. Extra marks were obtained if you mentioned the role of complementary distribution correctly, as explaining/describing why particular phonemes might receive quite different phonetic realizations (allophones) in different phonetic contexts.

This question was a chance for you to pick up more marks by picking basic terms from a range of areas in linguistics. If you had had problems with some of the other questions, this was an opportunity to demonstrate that you had engaged with the material for the course (and for Zwischenprüfung) more generally.

For marks for each of the pairs questioned, you should have said ideally what area of linguistics the terms were drawn from and then explained the difference between the items in the pairs as succinctly and technically as possible.

(a)    compositional semantics / lexical semantics is to do with a broadly logical decomposition of meaning, the first showing how the meaning of a whole syntactic unit (sentence, phrase, etc.) is put together out of the meaning of the parts and the second describing the meaning of the parts themselves, often words.

(b)   phonetics / phonology. Phonetics is the study of the concrete sounds of language, how they are produced, heard, etc. Phonology is the study of how those concrete sounds are organized into abstract systems, whereby particular languages group particular sounds together as not distinguishing meaning and others as the ‘same’ sound. The abstract sounds of phonology are called phonemes.

(c)    locutionary act / illocutionary act: drawn from speech act theory. The former the act of linguistic utterance itself; the latter the intended meaning of the act.

(d)   langue/parole: introduced by Ferdinand de Saussure to distinguish the abstract and idealized system of language itself (langue) from what happens to come out of people’s mouths when they speak, with mistakes, hesitations, re-starts, etc. Saussure (followed by Chomsky 50 years later) thought that only the former could give useful information about how language works; now, through conversation analysis, sociolinguistics, and others, we know better. Parole is also much more systematic than Saussure could know.

(e)    adjacency pair / side sequence: terms from Conversation Analysis. The most important notion to mention for the first term was that of prediction and what happens if that prediction is not fulfilled. Simply giving examples of adjacency pairs was a way to get some marks, but not all. A side sequence is then a sequence of turns that intervenes between the first part of an adjacency pair and the predicted second part of that adjacency pair. Simply stating that a side sequence was some off-topic additional talk was not enough as it is precisely its relationship to an adjacency pair that helps us define what it is.

(f)     verbal group / verb phrase: most of you managed to recall that the first of these is drawn from functional analysis, the latter from formal immediate constituency analysis. For full marks, however, you need also to say concrete how they differ. Some of you said that they were the same, but one if functional, the other formal. This was not good. For full marks you needed to say that the verbal group contains just the verbal elements while the verb phrase generally includes all those further constituents that fall under a VP. This is because the VP relates to the logical distinction made in philosophy and logic between a Subject and a Predicate, while the verbal group relates to Process in a Participants, Process, Circumstance analysis.

This question gave you a chance to say sensible things about dialects, about languages, about their speakers and cultures. While presenting examples of why the statement made could not be correct—for example, pointing out that the ‘dialect’ of the ‘language’ ‘German’ called ‘Bavarian’ might not be considered intelligible to speakers of ‘dialects’ of the ‘language’ ‘German’ spoken in the north of the country—obtained some marks, good marks could only be obtained by going further and presenting more useful characterizations of both dialect and language. Full marks required you to mention facts like the socio-political nature of deciding whether some group of language varieties is to be called a ‘language’ or not, and the existence of regional/geographic dialects in a continuum of language variation. Mentioning the four dimensions of language variation involved—time, regional, social, functional—got you some consolation marks but  were not directly relevant to answering the question unless used properly.

This question was also an opportunity to pick up some easy marks as the answers were mostly given on the exam sheet. Unfortunately, very few of you appeared to understand precisely what the question was asking and did much more work than was necessary. The question asked the following:

Given this tree, what are the phrase structure rules that it uses for NPs and VPs?

 Most of you then ‘proposed’ rules of varying complexity and accuracy. This was unnecessary, as if you have a tree in front of you, the rules are also completely and explicitly clear. You were not being asked to go further and make up rules which might or might not apply (as there are an infinite number of such rules that might have applied but did not). You should simply have looked at the tree and written down the rules that it required. There were not many of these.

For the NPs, for example, there are the following possibilities:

That is, 6 NPs. To find the rules, you of course simply look at what children these NP nodes have. And there are only two variations: one where the N′ is preceded by an article Art, and one where the N′ occurs by itself. The full and complete answer for the NPs was therefore:

• NP -> Art N′
• NP -> N′

You could, of course, put these two rules together, to give the single rule:

NP -> (Art) N′

No other rules for NP are given by this tree. And you can do this without even bothering to read the sentence!

The VP part of the question is even easier, as there are only two VPs in the entire tree:

These then give the rules:

• VP -> VP PP
• VP ->  V′ PP

Some of you tried to combine these two rules, which does not work. Any suggestion such as

            VP ->  (V′) (VP) PP

does not make any sense because it also means that it is possible to have both a V-bar and a VP in the same VP, which certainly does not occur in this tree!

No other rules were necessary for this question!! Indeed, many of the rules that were suggested, with complex combinations of articles, NPs, PPs, adjectives, etc., were simply wrong as they do not occur in the tree.

The participants, process and circumstances should also have been easy, although too many of you still made the mistake of stating that “with a friendly smile” is a circumstance. This is not a matter of interpretation: it is simply not possible for English sentences of this kind! Some of you noted that correspondence between participants and NPs: this was good and is what you should have done. But then it is also obvious that the entire NP “a kind old lady with a friendly smile” has to be the participant (and Subject, and Theme), not just some small part of this NP.

The other participant is found within the VP as usual and, in this case, more precisely within the V-bar.

The circumstances were (as is often the case), the PPs hanging of the VP nodes. There are therefore precisely two circumstances: “from a cage in the zoo” and “on Sunday”. Just stating that the circumstance(s) is/are “from a cage in the zoo on Sunday” was not enough, as it does not show that you have understood that there are two circumstances according to this tree. And, again, it is the tree that you should be looking at and analysing, not the sentence by itself.

Many of you made reasonable attempts at evaluating the VP rules. None of you really addressed the curious fact that we appear to have the same rule twice, although once with a child VP node and then again with a child V-bar node. Saying why this might be a sensible or not sensible thing to do would have received full marks.

For the final part of the question, people again often went off on their own and tried to apply their own proposed rules. This was not asked for and so was unnecessary. Nevertheless, if you proposed rules and then applied those rules correctly to this part of the question, you got marks for appropriately recognizing where and how structural ambiguity occurs in sentences—even though you were doing something slightly different to what was asked about.

What would have been the exact answer to this part of the question draws on the fact that, even with the three rules given above for NPs and VPs and which follow directly from the tree without any further assumptions or proposals, there could still be ambiguity about how the PPs can be related to the rest of the sentence. This follows simply from the fact that it is possible for a N-bar to be sufficient on its own or to have a PP too: i.e., from the rule (given by the tree and not invented) that:

            N′ -> N (PP)

So, for example, either the PP “in the zoo” is picked up as part of an N-bar (as done in the tree given) or it is picked up as part of a V-bar using one of the above rules. The difference can then be made visible by the other movement tests that become possible (you cannot move “in the zoo” if it is within an NP) and by the intonation (you pronounce the sentence differently depending on what you mean). Another possibility would be for the “from a cage” to be picked up inside the NP for the rabbits rather than leaving it as a PP for the VP (and therefore as a circumstance). There are several other possibilities too! So, stating what these ambiguities were received marks; relating them to exactly the correct rules received more marks still. Mentioning that this was a problem of PP-attachment was also rewarded.