Sei sulla pagina 1di 59

Sawek Mroczny Krl 2010. All rights reserved.

Just a good old yawn right in my face. Try not to do it. whenever I look at someone they start a huge yawn. It happens so often that people during lectures turn into vegetables. We are gathered here for the love of syntax. Romuald Gozdawa-Gobiowski, Office hours: Monday 1:00 2:00, room 10. Rgg.angli.uw.edu.pl Starting this Wednesday, we will be asked to log in to this site and select the middle choice. Then we follow some steps.

The key for this semester: syntaxrules,OK? Quizzes three times a semester. Timed answers give you extra points. An examination forum we will create our own questions. A normal forum - where we can exchange ideas, answer or ask questions. The points for the active participation in the lectures. What is language? There are two opposing schools of though. Generative and operative linguists Generative linguists they will not accept the definition of language along the line of its functions, it must be defined.

The building blocks that construct the language: 1. The sound system of the language the tiniest bits of language that we know are sounds. The sounds constitute the essence of our phonological competence, sounds that are at our disposal as native speakers. a. You know how to produce the sounds of your language, this is the simplest requirement. b. One can tell which sounds belong to the repertoire of the native tongue and which dont c. You recognize and unerringly produce all phonologically conditioned variants of the same sound; without knowing them. How it come that children learn all these subtle phonological differences. synu, nie wyaspirowae pocztkowego p, nie ma kolacji, adna matka go nie zaatakuje - the difference between Polish closed and open E as in ser, sie. why this conspiracy?

Sawek Mroczny Krl 2010. All rights reserved.

d. d. it is relevant for anyone to learn a foreign language to know this data Jak uczymy Jasia jzyka, dla kasy, to uczymy go bed, bad, bird dorzumy jeszcze bird, i Ja wymika. A critical period for language acquisition when for instance a child must be exposed to Chinese tones within the first months of his life to learn their phonemic significance. Jasiu, po polsku to jest pewna idiosynkrazja allofoniczna, ale po angielsku, zwyky dwik So, English e as in bed, occurs in Polish in pragmatically conditioned situation and in rapid speech. wart Wars paaca a paac paca? said quickly, enforces the English e Listen to Polish speakers powiedziaem, wiedziaem and in this context we find aesh. Just relate English sounds to their Polish equivalents, when explaining the pronunciation. 4. you know the rules for putting sounds together... A quiz: In which of these words is the G sound pronounced: Things, hanged, hunger, stronger, king, bingo? Engma in a Polish context: gga. Na samogosk mona krzycze: aaaa, a na P nie, ppp. In Polish, the engma is always followed by K. For English, G is not pronounced morpheme finally. Phonological competence depends on the early maturing macroneural circuits and must be acquired in the early childhood. If it is not, it will never be native like. Confer: the Henry Kissinger effect. He was German, he did not speak English when he went to USA, made a brilliant career, but still spoke an imperfect English, by which we mean its pronunciation. The phonology the earlier, the better. Sounds are put together to build morphemes. Morphemes produce syllables, then words. You have the ability to identify morphemes and to assign meaning/function to them. This is the morphological competence.

Sawek Mroczny Krl 2010. All rights reserved.

Morphological transparency when there is no surprising about the morphological level of its meaning Vs. Morphological idiosyncrasy something unexpected about the meaning of the word build of the same morphemes Teacher transparent Cooker idiosyncratic. Functionally idiosyncratic

Playboy, pijak. If playboy were transparent, then it should be a boy that plays. But there is some extra information imposed upon this word, it is semantically idiosyncratic. Pijak also idiosyncratic. On duo pije it does not mean that he drinks a lot of coffee. Semantic idiosyncrasy

Balon balonik, ogon ogonek There is no phonological rule to govern this transformation So, a structural idiosyncrasy. You have to learn them separately.

Pomidorek to jest may pomidor, selerek to jest may seler, but! ogrek, to jest ogrek.

Uwaga, piesi pay attention, stray dogs! Morphological idiosyncrasy.

Sawek Mroczny Krl 2010. All rights reserved.

This week the first quiz, so we should simply log in and try our chances. It is usually on for two days! The quiz contains four or three questions to answer with a single word or ABCD.

Morphological competence: You have the ability to identify morphemes and to assign meaning/function to them.

There is plenty of room down here for anyone who is short sighted.... or interested. Morphological idiosyncrasy: Antylopa, antychryst, antyciaa another example. Kawaler vs. kawalerka ka has the function of signalling the female performer of the action. so, kawaler kawalerka, it should mean the female spinster. But it does not. So, it is another case of morphological idiosynctasy.

3. Morphemes are arranged hierarchically within words. Syntactically. There is an internal composition to how words are build. Then, s is added. the core, er is attached. Teach er s

Resharpened

The first of March our lecture is moved somewhere else.

Sawek Mroczny Krl 2010. All rights reserved.

The ambiguity of the prefix UN UN attached can be to: verbs and adjectives: Unpack, undo Uncanny, unable, unintelligent, uninteresting, unsuccessful

Un + adjective: not Uninteresting = not interesting

Un+verb = do the opposite Unpack = do the opposite of packing

So, the conclusion is, that a 3-years old can distinguish between these two impossibly difficult cases, which cannot be done by grown-ups at the grammar course. There is no evidence that any child confuses these two. But children have no ability to reason in abstract terms. Abstract thinking comes at the age of 10.

What is the structure of uninteresting:

( to be found in the jpg file, that I will copypast here as soon as I can)

Unpacks:

Sawek Mroczny Krl 2010. All rights reserved.

Unlockable - it is ambiguous, it can be either not lockable, or one that could be unlocked. It is crucial for the preparation to the final exam to understand this ambiguity. For the ambiguity to exist, we need the possibility to add this un to the verb or adjective.

Folk morphology - the morphology of a simpleton, a small Ja or other creature of this sort. Naive views of native speakers can change the language forever.

Hamburg-er

ham-burger

Sawek Mroczny Krl 2010. All rights reserved.

Native English speakers felt that there was a ham in this hamburger. Normally, it was just a thing from Hamburg. So, something that was just a sequence of sounds, this burger, became a morpheme. It looked as though there were some ham, consequently, a new word is coined. Pullet surprise this is an attested error, made by twenty percent of college students who were taking a dictation. pullet surprise Pulitzer Prize.

Jzyczek uwagi uwaga z jzyczkiem? jzyczek u wagi, wagi szalkowej.

- child childru children - this form, children should not have happened. The productive plural in the Old English was ru. But then they forgot that this was already plural. Then, en as another plural died seconds after it was added. So children is as if double plural.

A whole nother story. other, another, an other Naranja (a norange) an orange the concept of norange coming from Spanish. Keep left kipilefti/vipilefti : Kirundi for roundabout(s). Architects of this country were trained in Great Britain. Kirundi marks singular and plural as a prefix. The prefix for singular is ki. The prefix for plural is vi. There were no roundabouts in Kirundi before. In GB, the sign before a roundabout states keep left. So, they thought it is the name for it... and kipilefti was coined. Folks thought that ki is a marker of the singular, and that the word is pilefti.

So, the language changes because of these naive preconceptions about the morphological structure of words.

Another example: James Gordon Bennett. He gave his name to a Cup Gordon Bennetts Cup And, in Polish: Prestiowe midzynarodowe zawody baloniarskie o Puchar Gordon Benetta. We have to know that the first name should be inflected.

Korona krlowej Anglii: The Queen of Englands Crown

Samochd mczyzny ktrego widziae a man that you saws car ?

Sawek Mroczny Krl 2010. All rights reserved.

- it looks as if you were adding s to seen but you are not, you are adding it to the Noun Phrase. The rule is that you add it to the phrase, not to the word. In English, genitive is syntactic, it attaches to phrases. It is morphological in Polish, though. Another rexample: my friend and Is decision.

Semantic (sub)competence

It deals with the dictionary meaning of words in your language and with sentence meaning as a function of the meaning of the words used in a given sentence. an informal meaning. The key word: dictionary meaning. Sentence meaning: as the sum of the meanings of the words in a sentence. As long as it is transparent, a sentence means what words in it mean. 1. You are able to perceive semantic ambiguity the first I saw her at the bank. the sentence is ambiguous, because of the two meanings of this bank the financial institution or the rivers. Mia piec, a nie mia wgla. Piec: as a noun and as a verb. And then: mia as a verb, or as a powdered coal.

Iraqui head seeks arms. Head as a top person or a part of a body.

Haemorrhoid victims turn to ice. They either transform into ice, or use it. The lamb is too hot to eat. it depends whether it is alive.

Hot too warm, sexually aroused, spicy. Popular?

Molestowaniem seksualnym Dariusza B. zaja si policja.

- Dariusz be, being sexually harassed by the police

Dwch zbliajcych si do czterdziestki braci przyrodnich dzieli prawie wszystko.

Sawek Mroczny Krl 2010. All rights reserved.

- they share everything, or they are different in everything? - Dwch either as a subject or an object.

Sawek Mroczny Krl 2010. All rights reserved.

Okay, so the quiz will be available for 24 hours. Semantic subcompetence It deals with the dictionary meaning of words in your language and with sentence meaning as a function of the meaning off the words used in a given sentence.... 1. You are able to perceive semantic ambiguity And, during the previous lecture, examples were provided.

2. You accept and actively propagate changes of meaning of individual lexical items. This relates to the concept of a lexical error. narrowing down the semantic fields it used to have. We propagate semantic change

Bielizna it is not as innocent as it looks. bielizna should be white. A hundred years ago it was white, today it is colourful, black etc. This is a change for which you are responsible. Zodziej it narrowed down its semantic field. Some 300 years ago it meant wrongdoer, and was an opposite of dobrodziej. So, zodziej was about anyone who did evil things. But today, this noun refers to just thieves. Akwen wodny pleonasm. Redundancy. akwen cannot be anything else but containing water. Mroczny Sawek, etc. But some military experts say akwen powietrzny. Wyjtek potwierdza regu an exception confirms the rule. It is a response to an argument providing an exception to the rule. there are exceptions to all rules. But what, the more exceptions, the better? `the exception that proves the rule, but then prove meant something else. The proverb is English, from fourteenth century. prove meant test, so it really means that an exception tests the rule, looking at it that way it makes more sense. A bigger half well, halves should be equal. But the bigger half is a shortcut meaning that one takes a somewhat bigger part of a thing. Semantically, this is an error. Someone who uses half does not use it in a mathematical sense. English authorities took courage to accept this, while Polish are more conservative. i po najmniejszej linii oporu it should be i po linii najmniejszego oporu. It shows you how native speakers behave. It does not have to be logical, but just make some sense in certain contexts. It is not interpretable semantically Tandem polsko-czesko-niemiecki tandem implies two parties, three parties is no longer a tandem.

3. You intuitively distinguish between homonymy and polysemy - Homonymy different (unrelated) lexemes have the same shape. It is an accident that bank of the river and the financial institution have the same form and shape. Polish: kopia (bro) vs. kopia (tekstu) yd karabin niesie or ydka rabin niesie

Sawek Mroczny Krl 2010. All rights reserved.

eyed bear - ??? - vs. Gladly, the cross Id bear Gladly, the cross-eyed Polysemy one lexeme has several meaning extensions there is a root and extensions of that core meaning. our inner semantic tools help us understand these subtleties. Head (body body part) vs head (of a department) Homonymy or polysemy? Date (=fruit) date (=calendar) - date (=sb you meet) Date- setting a date for a meeting, on a specific date. And then, date as a time of a meeting has grown, extended to mean someone you meet on that date Rodzina - as an example of chain polysemy. The second alteration includes something from the first, , but the third is only loosely related, or not related, with the first.

Numer buta

numer gazety

Numer = znak liczbowy Numer (pokj hotelowy) Numer (punkt programu)

In this case (above) each instance is equally distant. This is called radial polysemy. ysemy. You can have radial polysemy as a part of chain polysemy. niezy z niego numer, wykrci komu numer chain polysemy, but stemming from the radial polysemy of the base number

Sawek Mroczny Krl 2010. All rights reserved.

Ognisko muzyczne ^ Ogie ognisko ognisko soczewki


(an arrow pointing down)

Ognisko epidemii ognisko soczewki builds on focus. We are losing the fire part, what you are left with is the second part, that is added to ognisko soczewki, a lot of something in one place. Then, we preserve only a lot of sth in one place and derive other forms. 4. A crucial semantic skill we demonstrate is the recognition and use of metonymy. We do not even know that we use metonymy. Successful language use is metonymic. The conceptual relation X stands for Y

X is in a some way related to Y. More formally: One entity (the vehicle) is employed in order to identify another entity (the target) with which it is associated.

(a nurse, addressing another nurse) Woreczek ciowy poszed na USG. This is not just a man with a bile sack (woreczek ciowy). Dlaczego ta zakrzepica nie ley na naczyniwce it is depriving people of their humanity. (a waitress, addressing another waitress) Be careful, the ham sandwich has wandering hands dociekliwe, someone suggests here, sexual advances.

Typical conventional metonymies involve: a. Producer for product Pass me the Shakespeare on the top shelf here we mean Shakespeare as what he produced. Ive just bought a new Mercedes. youve just bought a car produced by the company, not the firm. Dostali nowe ksero. b. Place for event Iraq nearly cost Tony Blair the premiership. not the place, but the event, whatever happened there.

Sawek Mroczny Krl 2010. All rights reserved.

American people fear another Vietnam. not as a geographic region, but a repetition of what happened there. Uczynimy z Polski drug Irlandi. hopefully, it is a metonymy. c. Place for institution Downing Street refused comment. a location, but we mean a specific person working at a specific address in Downing Street Paris and Washington are in disagreement over the issue. Studenci protestowali, a Dziekanat milcza. d. Part for whole All hands on deck. She is a pretty face. To say ale z niego ... noga - an offensive way of refering to people, refusing to see them as whole persons, but just concentrating on a certain body part. Waciciel czterech kek. take the wheels away, there is no car. e. Whole for part My car has developed a mechanical fault. Here we admit our ignorance. Shes been bitten by a dog. But, the point is, what bit her, is the teeth. England beat Australia in the Rugby World Cup final.

There are two more major aspects of semantic competence (to be discussed later): 5. The use of metaphors 6. The use of formulaic language These concepts are related to syntax and pragmatics. And irony, what about that?

Syntax what we need to put words together to produce sentences. We want to say that syntax is indispensible for human language. Syntax is about the hierarchy of word arrangement in a sentence. From top to bottom, not a flat structure. Syntax is a weapon in communication. I gave her rat poison. I either gave poison to her, or to her rat.

Sawek Mroczny Krl 2010. All rights reserved.

22 March our lecturer is absent. And he does not know what to do with it.

So far: subcompetences of our linguistic competence. And we started on our syntactic competence: I gave her rat poison has two interpretations: One: she is dead, I poisoned her Two: the rat is dead, I poisoned her rat

The ambiguity must hide in the way these words are structured. It is not like I saw her at the bank where we have lexical ambiguity. In this case, the reason is syntax, since all words retain their meaning in both interpretations. So far, we have seen that the sequence of words Her rat - poison Is meaningful if arranged in two different ways, [her rat] [poison] or [her] [rat poison]

Consider the sentence: [I[ swallowed [her rat poison]]] So, we perceive syntactic ambiguity by building mentally tree diagrams like this. Syntax is about the hierarchical organization of words within bigger units (i.e. top-bottom arrangement ) rather than merely the linear (left-to-right) sequencing

You know the rules for putting words together to produce bigger chunks (units/phrases) This is the essence of your syntactic competence

1. This involves the ability to assign structure to sentences. If the same string of words can be assigned two different structures, the sentence is syntactically ambiguous. It has nothing to do with lexical or pragmatic ambiguity.

Sawek Mroczny Krl 2010. All rights reserved.

I shot an elephant in my pyjamas. another example. Both of them can be inferred from this sentence. Or: I heard that petrol can explode.

Do you see how syntax is instrumental in making sense of these jokes?

Archie (on seeing Edith come home with a bunch of flowers): Edith, where did you get all those flowers? Edith: Well, Arch, remember when I told you I was going to visit my friend in hospital? You said I should take her flowers. So I did.

The interpretation intended by Arch: take flowers for her, but her interpretation is that her flowers are a single chunk.

Ford: You should really prepare yourself for the jump into hyperspace. Its unpleasant like being drunk. Arthur: Whats so unpleasant about being drunk? Ford: Just ask any glass of water.

Or: Wife: My husband is getting a little queer to sleep with. Friend: What do you mean? Does he force you to indulge in unusual sex practices? Wife: No, he doesnt. And neither does the little queer.

Either a little queer is an adjectival phrase, or a little queer is a noun phrase, somebody with strange habits and this sb shares bed with them.

Sawek Mroczny Krl 2010. All rights reserved.

2. The second task: your syntactic competence involves the ability to detect synonymy.

The chicken crossed the road. The road was crossed by the chicken.

3. You perceive links holding between non-adjacent sentence elements.

Links, relations, that hold between word one and word seven. How do we know that these two words go together?

The sentence: buffalo buffalo buffalo. a group of buffalo harass another one. Buffalo buffalo buffalo buffalo. buffalo from Buffalo buffalo some buffalo, or those harassed are from Buffalo.

buffalo buffalo buffalo buffalo buffalo a group of Buffalo buffalo harass another group of Buffalo buffalo. Buffalo buffalo [whom] buffalo buffalo buffalo [] buffalo [some] buffalo buffalo.

My comment: it shows that the relative clause rule is recursive, and can be applied ad infinitum, creating endless sentences. Thus the above sentence can be expanded even further.

So ,the point is, that we add some relative clauses ad infinitum and still get the meaning of the sentence.

The teachers asked the students to behave themselves. the students + students The teachers promised the students to behave themselves teachers + students Links that hold between not adjacent sentence elements.

4. You recover missing words from certain grammatical strings

Will you join us for a drink?

Sawek Mroczny Krl 2010. All rights reserved.

I would if Tom did, but he cant, so I wont, either.

Notice how much is missing, but yet we understand it perfectly.

Tom left Peter to paint in Paris. We do not know who is in Paris. But this works only with left, and would not with told.

Finding the gun in the drawer worried us. How do we know that it was us who found the gun, and not the other party?

Finding the gun in the drawer he called the police.

5. Your very special gift, a mysterious link between syntax and semantics/lexicon: formulaic competence.

Formulaic competence: how come we can produce grammatical sentences at such a rate? Some scientists argued that the language should not be possible, because of its complexity, and two things people regularly do: grammaticality and the ability to understand it.

These are called two paradoxes of language. The formulaic competence, though, unlike all previous competences, is learned, not acquired. We are not genetically programmed to know formulas in a language.

Words strings which appear to be processed without recourse to their lowest level of composition are termed formulaic. formula is perceived as a single chunk, without worrying about its internal structure.

For instance: Silny or mocny? ___ herbata, ___wola, ___ w gbie, ____ Polska w ____ Europie

Sawek Mroczny Krl 2010. All rights reserved.

How do we know that the herbata can be only mocna, while wola is surely silna and not mocna etc. There is no obvious and easily accessible rule. You can only get it by listening to what native speakers say.

We can only learn these formulaic phrases one after another. So the people who learn their second language, will not acquire this formulaic competence easily. Formulaicity is not rule governed, it is accidental.

It is a measure of your native competence that you feel which word combinations are right. It is all about the preferred company for words we use and combine together. It is the power of collocation, just to know the part, to know what will follow.

Hence, foreigners are often perceived as unidiomatic since they avoid formulaic language or use it incorrectly. Sometimes the intuitions involved may be very subtle.

I eat breakfast at eight. - -- what is wrong with the sentence? Except that we have breakfast, not eat. It is grammatical, but not formulaic. Odesza w zielon dal. well, it was produced by a native speaker of English who is working at this institute. The reason we would not produce such sentence is that for us sina dal is more like a single word. Sprzedawa kotk w worku. why is this wrong? We know it as sprzedawa kota w worku, but someone who perceives this string as combined from individual words wil break it into small pieces, analyze them and make some mistakes.

I found conversing with you very pleasant, indeed Vs. Nice talking to you. The first is grammatical, the second is formulaic.

We won the match and the Italians lost, which was really surprising.

It is non-formulaic. We pulled off a surprise victory against the Italian team. This is a formulaic.

Sawek Mroczny Krl 2010. All rights reserved.

The Reds were victorious over the Whites. Formulaic as well. It shows that we, as the ones unlikely to produce these formulaic strings of words, are confined with rules, not ready to break them and come up with sth more idiomatic, imaginative.

What a pleasant event it is to see you. Tell me how your life is progressing at the moment. Vs. Hi, how are you doing?

Formulas act as individual lexical items. Thats why your formulaic competence can be seen as part of semantic (word-driven) competence. However, formulas also depend on syntax, (on morphosyntax) for their ultimate wellformedness.

The best way to see formulaic competence is as an interface between semantic and syntactic competence, overlapping with these two.

The psychological advantages of using formulaic language: Buys time for content planning (=aids the speakers production) Buts time for sentence processing (aids the hearers comprehension)

So, we just need to look at the meaning and ignore the form, which works without our own effort.

Formulas frequently undergo semantic contamination (esp. When the speaker lacks time or knowledge, under pressure, emotionally engaged)

1. My te odnielimy szkod. a breach, performed because of stress, presumably. 2. Jeden poprosi adnie, a drugi z ostrej rury daje. ( ostro daje + daje z grubej rury)

3. Ja mam tego pana serdecznie w dupie. Well, either serdecznie do or gboko w dupie. 4. Tak na zdrowy rozsdek. from the advert. na zdrowy rozum is the right one here.

Sawek Mroczny Krl 2010. All rights reserved.

5. W aypadku pasaerowie doznali powanych uszkodze 6. Opadw nie powinno by specjalnie sporo 7. Temu ukadowi trzeba zada klsk 8. Jasne jak drut! 9. Lepiej by ostronym ni dmucha na zimne 10. Chyli si nad grobem > ku upadkowi, sta nad grobem 11. Wszyscy podejrzani stan na awie oskaronych (przed sdem)

Formulaic competence, in a word, is an interface of syntactic competence.

Sawek Mroczny Krl 2010. All rights reserved.

The pragmatic competence. It is the system of beliefs each of us holds about the world, expectations about other people, cultural patterns shared by our speech community, all the non-linguistic factors that may have a bearing on utterance interpretation.

Speaker A: would you like coffee? Speaker B: coffee keeps me awake.

Etc..... to be honest, the latter part of the lecture further develops the concept of pragmatic competence, and then explores the notion of metaphor, exactly as it has been done during regular Intro. To Linguistics classes.

Sawek Mroczny Krl 2010. All rights reserved.

If Bill Clinton were the Titanic, the iceberg would sink.

Associations with the source domain: The Titanic, the iceberg, crash, the Titanic sinks Target domain associations: Bill Clinton and some unspecified force he will crash against Target domain extension: in the American context Clinton crashes against his political opponents Here, semantics and pragmatics go hand in hand, in order to extend the target, you need to have a pragmatic knowledge of who he was, and what he fought. Bill Clinton and his opponents collide, just like the Titanic and the iceberg do, but the result must be different because of the counterfactual if.

We are hungry for a victory. Phonological competence: correct pronunciation guide Morphological competence: inflection, derivation Syntactic competence: establishing relations holding between words within that sentence Semantic competence: establishing sentence meaning (theres something wed like to eat because we are hungry) Pragmatic competence: speaker meaning (victory is not sth we can eat, so relevance must be sought at a more abstract level) - We need to eat = we need victory - Everyone needs to eat = its only natural that we want to win - We cannot live without eating = victory is very important to us and well go to extreme lengths to achieve it In short: Desire is hunger

NB: in Polish DESIRE IS THIRST (spragniony zwycistwa, zaspokoi [] chu, ugasi dze)

Just as with formulaic language, metaphors tend to be contaminated, especially when the speaker is operating under pressure and/or is looking for a fresh, unused metaphor to impress his listeners Ta budowa to wrzd na zielonych pucach Polski - zielone puca is effective as long as it stands separately, but if we add another element, it becomes sth a little disgusting, green lungs get more literary. Komorowski stanowi ko niezgody w rozmowach koalicyjnych. Czy warto byo stawia rozmowy koalicyjne na szali tej koci. To say that a person is a bone of contention is bad enough. Czarnecki zebra du kup chopstwa.

Sawek Mroczny Krl 2010. All rights reserved.

- Polish used to have this kup, panowie, but the addition of this subtle zebra, to collect, distorts the picture and we tend to see another, underlying meaning of kupa. -- Antygona spluna moralnie na krla. -- Grafika trzyma poizom, chocia ju troszk czu [implies putrid breath or sth, not the intended meaning??] na niej zb czasu.

Semantic ill-formedness vs. Pragmatic unacceptability A sentence is semantically ill-formed (ungrammatical on semantic grounds) if its interpretation leads to a contradiction in terms.

A sentence is pragmatically unacceptable if tits interpretation calls for the suspension of our beliefs about the world A sentence which violates semantic criteria may still be pragmatically interpretable This triangle has four sides the definition of the triangle implies its having three sides, so we cannot even imagine a four sided triangle. A talking triangle is still a triangle, so this would be pragmatically ill-formed. But this example is semantically ill-formed. The bachelor next door is getting married again semantic ill-formedness, because bachelor is defined as someone who never married. My toothbrush is pregnant again. pragmatic this time. John is a living dead man semantic, because there is basic contradiction. Perhaps, though, a possibility of a zombu might happen to cross our mind. Speaker A: Art thou married? Speaker B: Well, I am, and I am not. pragmatic, probably?

A sentence may be pragmatically unacceptable even if it is well-formed syntactically and semantically. The boy fainted. The boy the girl loved fainted ??The boy the girl the teacher admired loved fainted. Well, theoretically possible to infer the meaning, but it is too complex.

Anyone who believes that the people who think that the earth is round are right is wrong If I truthfully utter this sentence, do I believe that the earth is round or that it is flat?. Flat. The cognitive load is too high. The earth is round = X People who think X are right = Y Anyone who believes Y is wrong

Sawek Mroczny Krl 2010. All rights reserved.

The soldiers marched past the school. The soldiers who marched past the school, fell. When we hear this sentence word by word, we cannot visualize it conceptually.

d Towards a theory of language acquisition GRAMMAR as a model of the competence of a fluent speaker We have to start looking at syntax by explaining how it is possible for a number of sentences to be infinite. Needless to say, the mechanism responsible for it must be grammar. I-language and E-language I-language - the set of rules and principles in the mind of a speaker specifying the set of sentences which zie could use if no non-linguistic factors were operative. (if zie did not have to die, or eat, etc)

Sawek Mroczny Krl 2010. All rights reserved.

E-language is an infinite set of all potential sentences of any natural language (all and only the grammatical ones) Alternatively: E-language is the set of sentences that some speaker could use if no non-linguistic factors were operative.

I-language is a finite set of rules operating on a finite set of words to produce an infinite number of sentences. But you cannot store all these possible sentences, as behaviourists believed. How can we demonstrate that the number of potential sentences of any natural language is infinite. The number of sentences is infinite if it can be shown that there is no limit to the number of words in an sentence (the longest sentence does not exist) Every sentence can be made longer by: a. Lexical means He met a beautiful girl. He met an intelligent, beautiful girl . He met an intelligent, beautiful, 24-years old French girl. - Theoretically, after some time, we can run out of words. He read an interesting book. He read a very interesting book. He read a very, very, very interesting book. - But this is cheating. b. Syntactic means - Conjoining (compounding) with and, or, but Tom read the report and wrote a letter but he forgot to sign it. Relativization sth that makes you human, makes your language function as a truly communicative system This is the book that Jack built. And the rule is recursive, remember? - This is the dog, that worried the cat, that ate the mouse, that ate the malt, that lay in the house that Jack built. Etc. c. Embedding - adding more embedded clauses to the main clause. His students love syntax. I know that his students love syntax. Peter thinks that I know that his students love syntax. Baej suspects that Peter thinks that I know that his students love syntax. -

Sawek Mroczny Krl 2010. All rights reserved.

That is the evidence, no longest sentence, = the number of potential sentences infinite (Elanguage) Therefore, we cannot have memorised an infinite set, so it all starts with the concept of rules (l-language) For the below: the argument is that small children learn the rules without being informed about them. How did we learn the rules of our native language? To answer this question we need to account for the following: Language acquisition is species-specific It is independent of general intelligence

Language acquisition: Rapid Subconscious Involuntary Equidistant (every child will learn any natural language with equal ease) Acquisition is complete before the child gains the ability to reason in abstract terms

Then, this leads us to the innateness hypothesis: Language acquisition is driven forward by an innate language faculty = genetic guidance on how to subconsciously analyze the language experience and devise a mental grammar of the language being acquired.

We are all equipped at birth with a set of rules which will ensure rapid, subconscious, involuntary L1A That set of innate mechanisms responsible for acquisition has come to be called UNIVERSAL GRAMMAR

A body organ, hiding in our brain

UG definitions: The essence of human language, the system of rules governing all human languages A genetic blueprint for language Knowledge in advance of input (expectations that the child has about the linguistic system before it is exposed to language)

Sawek Mroczny Krl 2010. All rights reserved.

Generative grammar is the only approach which assumes that the knowledge is language specific, a set of genes, otherwise coded information, only about language. The difference between cognitive and generative approach: the first will say that we learn language in much the same way as we learn chess, just an executive ability. Chomsky, though, says, that there is a more specialized device for language processing.

I f we are indeed generically aided in L1A, then it must necessarily be the case that UR-free acquisition is impossible.

Sawek Mroczny Krl 2010. All rights reserved.

UG and natural language acquisition practical considerations Universal Grammar -another definition: knowledge in advance of input (expectations that the child has about the linguistic system before it is exposed to language)

If we are indeed genetically aided in L1A, then it must necessarily be the case that UG-free acquisition is impossible. there must be things that we know about our language which cannot possibly have been deduced from the input, learned by analogy, etc. This is k nown as Platos problem or o Poverty-of-the stimulus argument (for the existence of UG) even though the input is imperfect, we end up with perfect linguistic competence The case: Christopher Taylor was diagnosed with brain damage at age six weeks. Although walking and talking were somewhat delayed, at about age 3 he showed interest in factual books telephone directories, dictionaries and books about foreign currencies and developed the ability to read not only in the usual fashion, but upside down or sideways as well. Most persons with savant syndrome have impoverished language skills as part of their basic disability, while musical, artistic, or mathematical skills flourish. Very rarely however, in an already rare condition, spectacular language (polyglot) skills represent the island of genius in stark contrast to other overall handicaps. (Darold Treffert, MD) Christopher Taylors case, in which he retained the ability to acquire languages. The way to test Taylors capacity is to expose him to a language violating the UG rules. For instance,, we automatically would reject a language in which negation would be formed by moving the third (any third) word in a sentence to sentence initial position. And the experiment was to see if Christopher would pick up such language. Neil Smith, British linguist, designed language, called EPUN, with a number of violations The girl - loves me Past Tense: me the girl loves. meaning: the girl loved me. Negation: loves - the girl me. Notice that you can do it, you do not believe that language like that is possible, but you can master that, memorize that. Translate into EPUN: The girl did not love me. Me loves the girl.

The control group consisted of twenty American college students.... they were given two tasks: one was Bask, and another EPUN. Twenty sentences from Epun and 20 from Bask, and they were to memorize these within an hour. The control group managed to memorize EPUN with equal excellence as they did Bask.

Sawek Mroczny Krl 2010. All rights reserved.

Christopher made 100% mistakes in all past tense sentences and 75% mistakes in the negative sentences in case of EPUN. Language is not a manifestation of and does not depend on intelligence. Christophers case shows how a person with an IQ in the mild to moderate range of disability can be a language genius. So far, it was the most convincing argument that the UG exists.

The reverse is also possible: the language faculty is impaired in the presence of normal intelligence. Stroke victims (who lose language ability but retain other intellectual functions) Aphasia (the most exciting disability) SLI (Specific Language Impairment) SLI children are unable to learn specific linguistic processes, although they understand their conceptual correlates. >> normal understanding of time and temporal relations but say things like: What did fell off the chair? What cat Mrs White stroked? Which coat was he weared?

UG involvement in L1A evidence from normal language growth How does the child figure out the rule responsible for the formation of questions involving the verb to be?

There is no evidence that any English child uses this rule Dolly mine can cry never becomes Mine dolly can cry?

English children do make mistakes in the process of question formation Is daddy should stay? There is virtually no input of this kind for children language acquisition, but yet one third of them come up with such things. Is dolly can cry?

Please note that some languages use the same process to ask well-formed questions.

Czy mog mie do nich zaufanie? Est-ce que je peux avoir confiance en eux?

Sawek Mroczny Krl 2010. All rights reserved.

Czy, est-ce que: examples of invariant (stable, fixed) question words that some languages possess ?

So, there must be a ban on wild grammars. Is Daddy should stay? Has the Polish-French structure: it uses IS as an invariant question word, i.e. it uses a question formation strategy that is attested in some world languages. Dolly my/mine can cry? Book the is over there? does not illustrate a question formation mechanism known to us. Christopher could not solve this type of puzzle, but he could do language.

Somehow the child knows what is possible and what is not. The child doesnt learn to form questions merely by listening to his parents. If he had, he would never produce questions like is daddy should stay

On the other hand, childs choice of question formation mechanisms must be restricted because there are some hypotheses that he never considers, e.g. children never attempt the move word-two hypothesis What is the relevant rule? Assume t he simplest possible formulation that could be acquired by analogy: Move BE to the beginning of the sentence. The book is on the shelf. Is the book on the shelf. Is the rule adequate? Sam is the cat that is black. English allows for two to be in a sentence. How a child knows which to move? So, maybe we need to reformulate the original rule: Move the first be that you encounter. Supposing that this rule works, and English obeys this rule...but it is wrong, because first implies counting, the ability to recognize which is first etc. the rule is immediately suspicious. It takes the rule out of the realm of linguistics. The cat that is black is Sam. the question: Is the cat that is black Sam? so, we have a question in which we (correctly) move the second to be.

Sawek Mroczny Krl 2010. All rights reserved.

Sentences involving two occurrences of BE constitute less than 0.5% in all caretakers speech corpora. In all likelihood, therefore, most children dont ever get to hear them. Yet by the age of 4 they are able to ask well-formed questions with one or two BE-forms in them. That is Platos problem: how to derive knowledge, if it is not in the data? The rule is: move the MAIN be? Which can be defined only in terms of structure, syntax. But the concept of the MAIN is not in the data. Child hears strings of words, but these are not tagged grammatically by the parents. Parents do not tag the verbs they use. They do not say things like: - Daddy main verb is adverbial compliment at work.

Sawek Mroczny Krl 2010. All rights reserved.

SVO English, Polish SOV Japanese OVB only some 1% of languages follow this pattern.

And there is a group of languages where it does not matter at all. Dyirbal (North Queensland, Australia)

Arabic Prime example of V S O There are more variations within each category, for example: subject before/after the auxiliary verb, before the verb but after an adverbial complement, etc. Here, we could argue that there is no UG, no common set of rules

Can there be a rule collapsing all these rules to a common patter? put the subject somewhere. is this rule true for all the languages discussed? it tells us that there is a subject in a sentence. There is just half of the rule. This rule is true only for languages which have the subject. Lets have a rule: put the subject in position X. Position X: a. Sentence initially b. sentence finally c. sentence medially d. pre-verbally e. postverbally f. anywhere

this variable can have a number of values. A parameter contains a variable (=an underspecified slot in the formulation of the rule). Selecting the right value for the variable is called parameter setting. Indentifying the right value for the variable...

Language acquisition is about parameter setting. The range of parametric options determines the limits of variability among natural languages. Natural languages will differ only respecting this ABCDEF placement. Well, so this is a huge difference. Parametric differences set a limit to the range of variability.

Sawek Mroczny Krl 2010. All rights reserved.

A more sophisticated example of a parameter: the Pro-drop parameter.

This model we formed should explain intervariability, but also intravariability within the same language, explaining the mistakes that some small native speakers make. a. You may drop the subject pronoun b. You mustnt drop the subject pronoun Does it matter whether we start with? [a] or [b]? no, it does not. What will happen if both Polish and English children begin language acquisition assuming the [+] setting of te Pro-drop parameter (=you may, if you want, drop the subject pronoun) Do I keep the subject, or do I drop the subject? Children in Britain ask themselves. The Polish child gets it right from the very beginning, but the English child must switch to the [-] Does the English child have any Positive linguistic EVIDENCE that his/her rule is wrong? There must be actual data, sentences, produced by his parents, telling him to reset the rule, telling that it is wrong. No, he doesnt. Every sentence with an overly realized subject is consistent with the assumption that you may drop the subject pronoun if you want to. No PLD would convince the child that he is getting it all wrong.

English children would have to rely on negative evidence. what would that be? Correction. This could be only of two kinds: a. Direct negative evidence = overt correction b. Indirect negative evidence (prolonged absence of a form, pattern, etc.) This can also count, technically, as a negative evidence.

Unfortunately, this would violate one of the basic assumptions about language acquisition: No-Negative Evidence Hypothesis The child attends to meaning and will remain insensitive to system corrections (e.g. morphosyntactic corrections) The child learns from positive examples of what people actually say rather than examples of what they dont say (children dont get any information about what is ungrammatical) There are hundreds of things adults do not say, but that are possible.

Is it possible that lack of evidence is an evidence of lack? it would be a dangerous assumption about the language.

Sawek Mroczny Krl 2010. All rights reserved.

An example of childs attending only to meaning: Child: Want other one spoon, Daddy. Father: You mean you want the other spoon. Child: Yes, I want other one spoon. Father: Can you say the other spoon?: Child: other... one... spoon. Father: say other Child: other Father: spoon Child: spoon Father: other spoon Child: other... spoon. Now give me other one spoon. So, the correction is futile, because the internal systems allows such Children do not react to system corrections They do not learn from what they dont hear. (does absence of evidence = evidence of absence)

The Pro-drop parameter cannot be set to the + value since the move from one value to the other must be accomplished by positive evidence (exposure to PLD) Negative evidence: absence or correction.

Starting with the [-] value. Now the English child accesses the correct hypothesis and the Polish child must switch. The Polish child can switch based on positive evidence. Can he manage that on positive evidence alone?

Yes, he can. Sooner or later the child is bound to hear a sentence without an over subject. When the relevant PLD accumulates (reaches some critical value) the child will know that he must reformulate his original mistaken assumption.

Sawek Mroczny Krl 2010. All rights reserved.

The move from a smaller to a bigger, more inclusive set can be accomplished by positive evidence. The reverse order (from the bigger to the smaller) would require negative evidence (noticing prolonged absence of a form or being sensitive to over correction) Children can be corrected in their pragmatic and lexical competence. Correcting childs mistakes is not a good practice, mistakes should be enjoyed in a specific way, in this process of language acquisition. If the languages differ at some points, lets introduce parameters/variables. And acquisition would be about picking them up.

There is a sacred word for all syntactitians: structure, hierarchy. ... the theory of constituents. Passivisation, relativisation: what does this rule do with constituents: it changes the order of constituents. A basic rule with no exception in all languages is structure dependence. Apart from parameters the grammatical system includes invariant principles Structure dependence Subjacency principle.

Structure dependence principle: all grammatical operations are structure dependent (=sensitive to the grammatical structure of the sentences they apply to) But this is a lie, we have to modify this rule.

Is there a rule that wouldnt be structure dependent, involving moving words, not constituents? For instance: attach sth to the third word in a sentence. Displacement of si in Polish. This can appear somewhere in the sentence regardless of the constituents. Tomek obrazi si na zielonych ludzikw. Tomek si obrazi.. Na ludzikw si Tomek obrazi..... Etc. Si must be attached to the first word. si jest pasoytem, it needs a word, not a structure.

MAIN is defined with ferenrence to grammatical categories and functions: following a NP which acts as a subject. ... Apart from parameters the grammatical system includes invariant principles Structure Dependence Subjacency principle

Sawek Mroczny Krl 2010. All rights reserved.

Range of possibilities for every unspecified slot defines the variability of language. Available in the Universal Grammar

Structure Dependence Principle All grammatical operations are structure dependent sensitive to the grammatical structure of the sentences they apply to) Recall question formation with BE: Move the MAIN BE Vs. Move the first /second BE

Inversion in question formation is defined with reference to the grammatical category of SUBJECT (is structure dependent Hence the contrast: Taxes will come down Will taxes come down? BUT: Down will come taxes *Will down come taxes. It all comes down to the fact that we are equipped with some rules that facilitate the process of language acquisition. Now: to the WORDS: Lexicon in CS studies What do native speakers know when they know words of their language Categorical information (N, V, A...) E.g. cat is a noun (+N) Szybko is an an adverb (+Adv) We know that cat is a noun, and we are able to explace it with other nouns, not necessarily giving the right name to this constituent. There are plenty of words in every language, and.. If we say szkoda e pani mnie nie sucha is it a noun? We cannot say st e marysia nie przyjdzie do szkoy because clauses are not introduced by nouns. szkoda seems to be one element category

English: He has enough time But:* he is politely student There is no other word in English with these properties.

Sawek Mroczny Krl 2010. All rights reserved.

So, the conclusion is, we are doing lexicon in terms of hundreds of categories, not five or six general. Categorical features these may be discourse dependent Common/proper -/+ count e.g. apple [+common, +count] What about Theres apple on your chin. a grandmother talking to her grandson, meaning a little bit, a mass. - There is cat on the bonnet Universal Shredder Principle - we put a countable noun in it, and we get a count version. c. Subcategorization frames some words determine the grammatical category of the words that follow them third property of our native speaker competence - When we think of behave it must be followed by an adverb. The knowledge of that is very intrinsic. e.g. convince: + [____NP] The verb convince subcategorizes for a complement belonging to the category NP Informally: what follows convince must be a noun phrase. Napada + [____PP] In this case specifying the category of the complement is not enough, as this would allow napada na konwj, po konwj, o konwoju Napada : V +[PP...na..]

Do the same for E: wait Wait: +[____PP] inadequate, as hthis would allow ungrammatical constructions Wait +[pp....for] Wait +[pp...on] d. Selectional restrictions pragmatic restrictions imposed on the choice of expressions which can occupy a given sentence position ** I have convinced. ** I have convinced about my mother I have convinced my mother ?I have convinced my cat * I have convinced my goldfish ** I have convinced my frying pan.

Sawek Mroczny Krl 2010. All rights reserved.

All these are in accordance with subcategorization, but there is sth wrong. Humanities studends have no problems understanding these, but on average,e this would not make much sense.

Understanding convince presupposes that the entity being convinced has reasoning powers. Frying pans are not known for their reasoning abilities, so we reject the sentence. it subcategorizes for an PN object but selects a rational object. Selectional restrictions are not syntactic in nature, they depend on our personal beliefs about the world Convince: V, +[____NP], <___rational> - one way of describing the relation, and rational is not linguistic, it is pragmatic. What can we say about Sobro and Asbro? Sobro killed Asbro accidentally? Asbro should be animate, and an adequate entity to kill, a significant one. Sobro a tornado, etc. Sobro murdered Asbro on purpose, violently. Sobro must be animate as well, human, perhaps. Sobro assassinated Asbro Asbro was important, VIP, and Sobro was commissioned to kill.

That is the fourth type of information on our disposal about words we know. When we are learning a foreign language, the data

e. Theta roles ( roles) /thematic roles/semantic roles specific semantic relationships holding between the verb and its arguments : experiencer, agent, object... I admire my teacher experience I drink milk agent The list of theta roles: Agent instigator of an action, one who intentionally initiates it Patient person or thing undergoing th effect of some action Recipient entity receiving something but not actively involved in getting it INSTRUMENT inanimate cause, means by which something comes about EXPERIENCER entity (usu. Animate) experiencing some psychological state

The FBI arrested Larry Boysterous. The FBI [agent] arrested Larry Boysterous [patient] The wind [instrument] blew off the car [the patient]

Sawek Mroczny Krl 2010. All rights reserved.

The prisoner died. Lets consider the psychological process that is taking place within us. How does this influence our grammar?

the noun phrase: the death of the prisoner of is associated with the experiencers The prisoner protested. The protest by the prisoner by is associated with agents. a book BY someone. Etc. a book of lies book with lies within.

What about: Nie pacz MI tutaj. Nie umieraj mi tutaj. E: dont die on me. it is unfortunately ambiguous

mi the role of an experiencer is forced here. I experience discomfort because of your crying. Polish allows such an forced experiencer for every verb nie ucz mi si tutaj lekarka zamaa mi rk meaning that the doctor broke her hand, not mine. fascinating. si another fascinating bit. It deprives the verb of one of its theta roles. Tomek gotuje zup Zupa si gotuje Si is sending a signal, a theta role I want to avoid mentioning. In standard Polish, the theta role that gets eliminated the role of the agent. What children do is the extending of it, to cover patients as well. : prosz pani, on si bije jak on si bije, to Ciebie nie boli . on the one hand, we have forced theta role, on the other: dethematized theta constructions. Personally goes with experiencers. Deliberately goes with agents. I hate personally, I hit deliberately.

The criterion: Clause A: Each argument is assigned one and only one role. Clause: B: each role is assigned to one and only one argument. Tenses.

Progressive tense compatible with agents, and not with experiencers. Agenthood presupposes your control of the situation. Examples: I am listening things vs. I am hearing things. I am seeing things vs. I am looking at things

Sawek Mroczny Krl 2010. All rights reserved.

Progressive is really about change. growth, etc, and then we can use experiencers to describe the action. I am owning a car pointless. But today people are owning more and more cars this is okay, a change is described. The criterion: Clause A: Each argument is assigned one and only one role. Clause: B: each role is assigned to one and only one argument.

(copied here) There is a great difference between these two.

Mary loves Peter 1 i 2 J

The above: theta role grid. The first line: theta roles The second line: arguments

Use the theta criterion to explain the ungrammaticality of the sentences below: 1. John kissed Mary the bed. 2. They offered to the best candidate Theta grid represents the compliments of the verb, it represents what the verb means. Arguments necessary for the sentence to be grammatical.

Based on what you know about theta roles account for the ambiguity of the sentences below. John felt a real idiot experiencer or agent. Mary opened the door with John In the second sentence John is either a co-agent, or an instrument.

Sawek Mroczny Krl 2010. All rights reserved.

Please put down three sentences. The door opened. there is only one theta role and argument in the grid. Mary opened. Mary opened the door.

Assume that one of these is ungrammatical. The second, because the theta argument is not assigned to the theta role that calls for it.

Sawek Mroczny Krl 2010. All rights reserved.

Passivisation A generative UG-based account

Tom will be arrested.

The most common tree diagram is not satisfying, because...

In this sentence, Tom is assigned a theta role of patient/theme. Normally, the verb sits next to Tom and assigns him the role a patient, in a sentence like The police will arrest Tom. Evidence against such a representation: a. Subcategorisation transitive vs. intransitive arrest??? In the representation given, it looks like arrest is intransitive, whereas the meaning is clearly transitive. b. Idiom chunks (class of NPs highly restricted in their distribution, used only AFTER some specific verbs) for example: zjeczae maso rancid butter Keep tabs Pay heed Pay homage Tabs, heed, homage dont have the distributional freedom enjoyed by other nouns. They are always connected with specific verbs. Cf. Heed vs. attention Can I have your attention / * heed We need to pay more attention / (=heed) to his words

There is one context where heed doesnt follow the verb pay Little heed was paid to her proposal Close tabs were kept on all syntax lovers Due homage was paid to the dead. The argument: is it not that structurally the noun still follows the verb, and so it is not an exception?

Movement analysis Only one verb to arrest, with the same subcategorisation properties both in the passive and in the active

Sawek Mroczny Krl 2010. All rights reserved.

In idiom chunks...

The Deep Structure It determines the basic, underlying, semantic relations, holding between the elements of a sentence. Here, Tom is a patient, and the only way to capture that in Generative Syntax is to say that Tom is a sister of VP.

The Surface Structure it shows the sequence of pronouncing things. The trace is a signal that there is something there, and we cannot formulate a sentence like Tom will be arrested Peter necessary to prevent the generation of such sequences.

Passivisation Rule Move the immediately postverbal NP argument to the subject position

This rule handles a whole range of cases both in Polish and English. Eg: Omwi project Mwi o projekcie He kicked the ball Projekt zosta omwiony. *** the ball was kicked ***

He waited at the bus stop

*** - the problem is here that the NP is a sister of PP.

However, there are cases where the rule OVERGENERALIZES and other cases where it seems to UNDERGENERALIZE. Overgeneralization the rule is too general, allows cases which are in fact ungrammatical Undergeneralization the rule is too restrictive, rules out cases which are in fact acceptable.

Sawek Mroczny Krl 2010. All rights reserved.

Overgeneralization in Polish: Podzikowa MArysi *** Ciotka handluje warzywami *** Boj si choroby *** Mucha przeleciaa Marysi koo nosa

***

Overgeneralization continued - Preserve the general V-NP rule of passivization but add a language-specific rule .

Undergeneralization in English Tom looked at the chimney -- The chimney was looked at. They talked about the project The project was talked about

How to account for these sentences, at the same time preserving the general rule that only the immediately post-verbal objects passivise? Why are we fighting for this rule? Because it is so simple. We want rules to be elegant.

Look at the chimney Talk about the sex

What evidence do we have to prove that look at is a prepositional verb? Constituency test If a string of words (e.g. at the chimney) is a constituent (a complete chunk) then things will happen to all of these words at the same time if something happens.

Sawek Mroczny Krl 2010. All rights reserved.

Movement test You may want to move a chunk somewhere else within the sentence, for example to the very front. If the resulting sequence is grammatical, what you moved is a constituent. They looked at the chimney At the chimney they looked.

The chimney was looked at Vs. *** At the chimney was looked If the verb look is prepositional, the second sentence should be possible.

A: Wast the picture looked at? B: *** No, at the chimney [was] it seems that here at the chimney is no longer a constituent.

Adverb insertion: They looked at the chimney They looked discreetly at the chimney. But: The chimney was looked at The chimney was looked discreetly at. We again see that in the passive voice, this supposedly prepositional verb acts like a phrasal verb. If our tests point out that there is a pattern, there must be something to it. We have to change our interpretation: The verb look at is prepositional in the active. In the passive voice it behaves more like a phrasal verb. The process needed is known as RESTRUCTURING [V-PP] [V c NP] - (small c in the lower index standing for complex verb)

Sawek Mroczny Krl 2010. All rights reserved.

Is Restructuring in any way restricted? If it is not, every English verb followed by a PP should have a passive equivalent. They talked about the project. They talked in the Oval Room The project was talked about. The Oval Room was talked in.

It has to be restricted in a way that is psychologically learnable, to reflect a pattern in reality that we are familiar with.

A more revealing contrast: Many people lived in this village. Many people died in this village. . This village was lived in by many people. * This village was died in by many people.

Restructuring is limited to cases where the V+P sequences that represent single semantic concepts, replaceable with single words. [Talk about] the project = discuss the project

[look at] the chimney = watch the chimney [live in] the village = inhabit the village [die in] the village = *** the village

Examine the possibility of passivising the underlined NPs within the sentence. They took advantage of the situation.

The situation was taken advantage of. Or: advantage was taken of the situation. Think about this sentence at home..

Sawek Mroczny Krl 2010. All rights reserved.

Passivization semantic conditioning Restructuring change in the syntactic properties. .

Restructuring is limited to cases of V+P sequences that represent single semantic concepts, replaceable with single words

Sala 400 16.00 jutro. The final quiz is said to be available from today night onwards.

[talk about] the project = discuss the project

In Polish, possibly only Accusative NPs can be passivized.

They took advantage of the situation... The situation was taken advantage of The advantage was taken of the situation ;) The situation was exploited. - Restructuring

[take advantage of] exploit (the situation)

Passivization pragmatic conditioning

Passivization is pragmatically conditioned: the likelihood of a passive sente nce being used increases if the NP object is perceived as sufficiently affected by the action. this case is more subjective, and the opinion will vary across the sample of Native Speakers.

Compare: Many generals deserted the army. :-- the army was deserted by the generals.

Sawek Mroczny Krl 2010. All rights reserved.

But: Private Smith deserted the army. *The army was deserted by Private Smith. This illustrates the concept of ascribing importance to the words, whether we can formulate a sentence or not depends on our pragmatic competence.

The same applies to restructuring. It will be possible n pragmatic grounds (the speakers judgement about the degree of affectedness) even if the V+P sequence is not a coherent semantic unit.

Paco slept in this bed. Is passivization possible here? This bed was slept in by Paco No, it is not. It is ungrammatical (or at least unsemantic)

He seemed undistracted by finding out that our bed was slept in by Queen Victoria the main memory of the events .. (the same bed was slept in by his pursuer) Now we realize that this impossibility to passivize PAco is more about pragmatics. It is the bed that changes. The queen, having slept in this bed, transforms it into a museum piece. And then it can be passivized.

Another case: To be lived under? Have you looked under that rickety bridge? It is disgusting there. It was lived under by a troll some time ago.

Under normal circumstances, such sentences would be unacceptable, because they require pragmatic context normally unheard of. the door was stood behind by Susan Boyle

Sawek Mroczny Krl 2010. All rights reserved.

Undergeneralization Cd. They gave the book to Mary The book was given to Mary. Mary was given the book [to]

[give thebook to] is not a likely candidate for restructuring, so Mary was given the book to Should be ungrammatical. mary was given the book derives from they gave Mary the book -

Both agree then with the rule that the immediately following NP can be passivized. A related situation: John gave Mary the book Mary was given the book The book was given Mary.

A prediction: There will be as many passives as there are actives. In each case only the immediately postverbal NP will passivize.

Consider: I delivered the parcel to the woman * I delivered the wmoan the parcel. So, only one passive should be possible. So: The parcel was delivered to the woman. *The woman was delivered the parcel. But- in North of Britain both forms are possible.

Sawek Mroczny Krl 2010. All rights reserved.

Why does passivization involve movement? Why doesnt English allow sentences along the lines of It was arrested Tom There was given the book to Mary

Recall Strict Adjacency Principle. love syntax passionately vs. *love passionately syntax.

Love Mary vs. look at Mary - here, at has the same role as love in the previous one. A relation between the verb and the noun is the same as between the preposition and the noun. This relation cannot be broken, once one is established.

Every noun in every natural language must have case assigned to it. Typical lexical case assigners (cross linguistically) : V and P. In positional languages Vs and Ps can assign case to their immediate right-hand (left-hand) sisters. In languages such as Polish this closeness is not so much required.

dny wadzy the case in Polish where case assignment is performed by an adjective.

In every language there is also a set of functional case assigners. To love syntax is well formed as a VP John to love syntax is ill formed as a sentence. Why? It requires a suffix. The answer lies in inflexional affixes.

The verb must be tensed oczasowiony Every sentence, must have a tense.

Sawek Mroczny Krl 2010. All rights reserved.

A tense is non-lexical, functional case assignment. Every sentence must have INFLECTION (=TN) Why? Because INFL is a functional case assigner. In English it assigns [+NOM]... to the left.

John told Mary about the accident..... here the task is: mark with arrows which assigns case to what.

What does this have to do with passivizaiton?

The passive affix deprives the verb of its ability to assign case it makes the verb morphosyntactically passive. On the one hand of a sentence, you have an empty slot to which a case has been assigned, on the other: a noun to which no case can be assigned, because the verb is passive.

Inflection and movement two flaws on the perfect body of language. Chomsky, 1960/70

Sawek Mroczny Krl 2010. All rights reserved.

The UG model of L1A first language acquisition. The consideration: what is it in English that drives the movement of the NP in passivisation? Weve shown that this is concerned with the case assignment.

It seems that the model of L1A is very much involved with passivisation. One of the main questions of Syntax is the L1A, how it happens? We have the experience of language around you... and what you hear becomes an input of the magic black box of UG and what comes out is the grammar (I-language), a live set of rules, not book grammar. We add another dimension to it, we say that exposure to language is not enough, it needs to be interpreted. And this is the job of UG. Two senses of grammar... Grammar of X I-language = NS competence of his/her language, it makes you native speaker of your language Grammar is also the study of that competence = MODEL of the competence (linguistic knowledge) of a fluent NS. Every grammar book is in a sense trying to model what native speakers know. But most attempt to be theoretical, concerned with intuitions, not a model. Features of grammar as a competence model There will be three criterions in evaluating the model - Levels of adequacy: Adequacy: - match between a sentence and its representation. 1. Observational adequacy: the strings generated by the grammar match all the actual sentences and only actual sentences. A grammar is observationally adequate if it can handle the PLD (primary linguistic data) the input that UG works upon, adult speech as heard by the child. - A representation of every individual sentence, and if all are well-formed, then the model meets the requirement of the first criterion of adequacy. Thats the basic level. But in case of three years olds three word utterances like john must eat and must he cry? would suggest a rule move the second word to the beginning of a sentence. But we know that this would not agree with our intuitions, the second level. Slightly more sophisticated.... : 2. Descriptive adequacy - First of all, it will satisfy observational adequacy. - Then, the structural representations (e.g. tree diagrams) correctly represent the speakers intuitions about sentences. A grammar is descriptively adequate if it can handle the facts of language. - In other words, a descriptively adequate grammar handles the linguistic competence that comes out of the UG. And, finally... the most sophisticated level of adequacy... 3. Explanatory adequacy - An explanatorily adequate grammar links PLD and linguistic competence, it accounts for the mystery of the acquisition process, i.e. it must provide a principled reason why we know what we know about our language given the obvious imperfections of the input data. It gives your rules which are learnable, compatible with what we know about the workings of the human mind. Some rules may be observationally and descriptively adequate, describing correctly the human behaviour; but the problem is that they usually presuppose too much about how our mind works. We -

Sawek Mroczny Krl 2010. All rights reserved.

need rules that are psychologically real, short and at the same time satisfying. They should make accurate predictions about what counts as data. Predictions about human cognition should be also possible. The limitations on the input data may pose a problem for rules that fail at the level of explanatory adequacy. A linguistic example. - How many tenses do we have in English? Some other questions this may trigger: what do you mean, what do you recognize as a tense, etc. But a good question: how many of the m are actually used in every-day discourse. From a pedagogical point of view, there may be at least sixteen possible tenses, eight for present+future, and eight for all sorts of past + unreal past. Sixteen basic tenses and modal + verb variations (present, past, perfect): I can go, I could go, I could have gone.. that gives at least another fourteen combinations, in sum over thirty verb forms each coming in all modes (as questions, negations etc.) To handle this we need over ninety rules of FORMATION. These have to be accompanied by the relevant RULES of USE. Assuming only three separate uses of each verb form, this yields another ninety rules. So, there are 180-220 rules to describe a minor area of English grammar ?! But a child does not hear many of these, and yet is competent to use them when the need arises. Its a Martian account of the English tense system. Something is going wrong here. Even if this account satisfies observational adequacy or descriptive adequacy it abysmally fails at the level of explanatory adequacy. An alternative account then must be formed... - That would predict automatically almost question formation and negative sentence formation... and would reduce the number of tenses to TWO. Before we start, a comment on Martian like explanatory adequate model of pedagogical rules... we learn a system that is flawed, and then we teach it. why? So: we want to reduce rules of formation from 16 to two. Every sentence consists of a subject NP, auxiliary element (AUX) and a VP. S

NP

AUX VP TN -PRES/+PRES go

Affix Hopping (-PRES + go = went) The AUX node is the simplest one. We often put there time aspect. -/+ PRES is an affix that needs to attach to the verb. Not just any verb but the first verb to its right. This is the essence of the AFFIX HOPPING TRANSFORMATION. This accounts for two tenses.

Sawek Mroczny Krl 2010. All rights reserved.

- I go - I went. What about I can go, I could go ? So, the AUX node can optionally include modality. AUX TN MOD -PRES

Lets draw a diagram of I could go S NP I TN -PRES AUX M CAN VP V go

In this case, AFFiX hopping still works, it jumps to can.

Another thing we may need to account for perfect tenses is ASPECT. AUX Tn -PRES ASP(ect) PERF have en

So, it consists of two things (besides the optional modal). Now, identify the morphemes that signal this aspect? have and ed/ - or, as marked on the tree, it is have and en, the second one is the general marker of 3rd form.

Now, draw sth for I have gone

Sawek Mroczny Krl 2010. All rights reserved.

Now, we have to think about .... I am going. AUX Tn ASP

Here, by the same token, +PRES will jump to be and, -ing will jump to go.

Sawek Mroczny Krl 2010. All rights reserved.

Now, we have to combine these two, by looking at things that combine them. She had been writing. Hint: determine the correct order of PERF and PROGR by applying Affix Hopping.

To produce a sentence She had been writing

Now, a modal. She could have been writing.

Sawek Mroczny Krl 2010. All rights reserved.

How to tie it up with passivisation?

Lets add the PASSIVE node. Passivisation is about depriving the verb of its morphological property of case assignment.

Sawek Mroczny Krl 2010. All rights reserved.

Tom [threw a fit], Tom [threw] a ball. Here we conclude that the verb with the object, the VP, assign a theta role to the subject. Here, in the first one, the role assigned is that of an experiencer. In the second, it is agent.

Another example: Lydia took cover, vs Lydia Took fright.

The Auxiliary... AUX TN Tn (M) (ASP) +/- PRES

ASP -> {PERF, PROG} or both. PERF PROGR PASS have en be ing be en

Tom would have been followed. Draw a tree diagram of that ...

The structure of AUX will from the left to the right include things as they are listed above from top to bottom. So, PASS nod will be the rightmost, while Tn will be the leftmost.

Tense it is a functional case assigner. Functional case assigners assign the case to the left, not to the right, as the lexical means.

Now, try and draw.... the detailed tree diagram for the sentence: That job should have been being offered to Barbara.

* It will be offered a job to Barbara. - now, our job is to explain why this sentence is ungrammatical.

Sawek Mroczny Krl 2010. All rights reserved.

a job cannot jump to the initial position, because that slot is occupied by it. then, a job remains without a case assigned. And so sentence is ungrammatical. You would say that there is an uncased marked noun phrase, for which there is no cased position available in the sentence, because the only potential slot has been filled by a dummy, a semantically empty element it.

Barbara was being offered a job to. What happens here is that Barbara is quicker that the job, and it is Barbara who lands in the subject position, leaving a job without a case assigned. Barbara blocks the only elicit movement. Moreover, Barbara is assigned two cases, by to, and by the tense. It is too much.

Now, Shakespeare. We will talk about Shakespeare.

Did you went is ungrammatical, because there is tense assigned twice. Well, too much!

In Shakespeare, we have two ways of forming questions: 1. Knowst thou not? Sees he that it is so? 2. Didst thou not know? another way. So, there was already a transition.

Is there a pattern of asking these questions? Can you stay? Did he see you? Knows he not thy voice?

It is: -

To ask a question move TENSE to the front (pre-subject) position. But, some verbs became lazy and did not want to move. Saw you my daughter see decided not to move. At that point, English grammar had to invent a word that would carry the tense to the front.

The emergence of do saved the question formation.

Potrebbero piacerti anche