Linguistics jegyzet (német egyetemről, de angol nyelvű)

Országok listájaHungarySzegedi TudományegyetemBölcsészettudományi KarAnglisztikaIntroduction To LinguisticsJegyzetekLinguistics jegyzet (német egyetemről, de angol nyelvű)

EINFÜHRUNG IN DIE SYSTEMTHEORETISCHE SPRACHWISSENSCHAFT

Institut für Anglistik und Amerikanistik Humboldt-Universität zu Berlin

Prof. Dr. Susan Olsen WS 2004/05

Table of Contents 0. General Organization ....................................................................... 2 1. Introduction to Linguistics .............................................................. 4 2. Morphology .................................................................................... 9 3. Syntax ............................................................................................ 18 4. Semantics ....................................................................................... 29 5. Pragmatics ...................................................................................... 37 6. Phonology ...................................................................................... 45 7. Neurolinguistics............................................................... 53 8. Language Acquisition ...................................................................... 59

2

0 0.1

General Organization of Lecture Purpose of Lecture

Diese Vorlesung vermittelt einen Einblick in Fragestellungen, Methoden und Ergebnisse der modernen Grammatiktheorie. Am Beispiel des Englischen werden die verschiedenen Kernbereiche der Grammatik - Phonologie, Morphologie, Syntax, Semantik, Pragmatik vorgestellt und in ihrem Zusammenwirken beschrieben. Zusätzliches Augenmerk gilt dabei der psychologischen Basis des Wissenssystems 'Sprache' (v.a. dem Erwerb und der neurologischen Grundlage dieses Wissens (= Erstspracherwerb und Neurolinguistik)). Vorlesungssprache ist Englisch. 0.2 0. 1. 1.1 1.2 1.3 1.4 1.5 1.6 1.7 2. 2.1 2.2 2.3 2.4 3. 3.1 3.2 3.3 3.4 4. 4.1 4.2 4.3 4.4 5. 5.1 5.2 5.3 6. 6.1 6.2 6.3 6.4 6.5 6.6 6.7 7. 8. Program of Lecture General Organization of Lecture Introduction to Linguistics Linguistics Rule System Universal grammar Language Universals Cognitive Science Animal Communication References Introduction to Morphology Structure of Words (Inflection vs. Word Formation) Compounding Derivation by Suffixation References Introduction to Syntax Constituent Structure Complementation and Modification The Verb Complex References Introduction to Semantics Meaning Sense Relations Predicates and Arguments References Introduction to Pragmatics Speech Acts Conversational Implicatures References Introduction to Phonology Evolution of the Vocal Tract The Phonological Base: Phonemes Articulation of Consonants Articulation of Vowels Structure of the Syllable Phonological Processes References Neurolinguistics: Structure of the Brain, Aphasia Language Acquisition

3

0.3 Recommended Literature 0.3.1 Introductory Texts Akmajian, Adrian, R.Demers, A.Farmer & R.Harnish, 1995. Linguistics. An Introduction to Language and Communication. 4th Edition. Cambridge, Mass.: MIT Press. Aitchison, Jean, 1995. Linguistics. An Introduction. 4th Edition. London: Hodder & Stoughton. Erickson, Jon & M. Gymnich,1998. Grundkurs Anglistische Sprachwissenschaft. Stuttgart: Klett. Fromkin, Viktoria (ed.), 2000. Linguistics. An Introduction to Linguistic Theory. Oxford: Blackwell. Fromkin, Victoria & R.Rodman, 1998. An Introduction to Language. 6th Edition. New York: Harcourt Brace College Publishers. Kortmann, Bernd, 1999. Linguistik: Essentials. Anglistik-Amerikanistik. Berlin: Cornelsen. Meyer, Paul Georg et al., 2002. Synchronic English Linguistics. An Introduction. Tübingen: Narr. O'Grady, William, J. Archibald, M.Aronoff, J. Rees-Miller, 2001. Contemporary Linguistics. An Introduction. Third Edition. New York: St. Martin's Press. Radford, Andrew, M. Atkinson, D. Britain, H. Clashen & A. Spencer, 1999. Linguistics. An Introduction. Cambridge University Press Yule, George, 1996. The Study of Language. 2nd Edition. Cambridge University Press. 0.3.2 Lexica

Bußmann, Hadumod, 1990. Lexikon der Sprachwissenschaft. 2nd Edition. Stuttgart: Kröner. Bußmann, Hadumod, 1996. Dictionary of Language and Linguistics. Translated and edited by Gregory Trauth & Kerstin Kazzazi. New York: Routledge. Glück, Helmut (ed.), 1993. Lexikon zur Sprachwissenschaft. Stuttgart: Metzler. 0.3.3 Further Study

Plag, Ingo, 2003. Word-Formation in English. Cambridge: Cambridge University Press. Burton-Roberts, Noel, 1997. Analysing Sentences. London: Longman. Crystal, David, 1995. The Cambridge Encyclopedia of Language. Cambridge University Press. Crystal, David, 1997. The Cambridge Encyclopedia of The English Language. Cambridge University Press. Kempson, Ruth, 1977. Semantic Theory. Cambridge University Press. Newmeyer, Frederick (ed.), 1988. Linguistics: The Cambridge Survey. 4 vols. Cambridge University Press. Osherson, Daniel (ed.), 1995. Language. Vol. 1 Language, Second Edition. Cambridge, Mass.: MIT Press. Radford, Andrew, 2004. English Syntax. Cambridge: Cambridge University Press. Saeed, John, 1998. Semantics. Oxford: Blackwell. Spencer, Andrew, 1991. Morphology Theory. Oxford: Blackwell. Spencer, Andrew, 1996. Phonology. Oxford: Blackwell. Stillings, Neil, et al. (eds.), 1995. Cognitive Science. An Introduction.Cambridge, Mass: MIT Press.

4

1 Introduction to Linguistics 1.1 Linguistics Linguistics is concerned with the nature of language. It is the scientific study of natural language, i.e. language that is acquired and used by humans. The structure of human language is highly complex, a fact that is not always apparent to speakers because of the ease with which is it learned and put to use in daily communication. Language is a domain of human knowledge; it is an abstract system made up of different structural levels. These levels are phonology, morphology, syntax and semantics, which provide the discrete units that make up the respective level (phonemes, morphemes, phrases, lexemes) and the rules and principles that determine the way these structural units are combined and ordered. The rules capture the patterns among the units at each grammatical level. These levels, or components of grammar, are the subfields that formed the core of structural linguistics, developed by the Prague school in the 1920s, and also make up the central areas of generative grammar and its offshoots (lexical functional grammar, head-driven phrase structure grammar, cognitive grammar, optimality theory, etc.) that have been the dominant theories of grammar ever since. 1.2 Rule system The most fundamental assumption about the structure of language is that it is rulegoverned. Rules govern the combination of units at each linguistic level. However, linguistic knowledge is subconscious knowledge. It is learned unconsciously with no awareness that its patterns are being internalized. The speakers of a language are as unaware of the rules of their language as they are of the rules that underlie their other cognitive abilities. Nevertheless, the ability to speak and understand a language reveals knowledge of a rule system that is quite complex. The notion of linguistic rule has a special sense. Linguists do not study prescriptive rules such as `Never end a sentence with a preposition.' (e.g., This is the chair he sat in.). Rather, they are concerned with formulating the descriptive rules that express the generalizations that describe a speaker's knowledge of his language, his linguistic competence. In investigating linguistic competence, linguists focus on the mental system that underlies the speaker's ability to produce and comprehend linguistic expressions. This system of knowledge is called a grammar. The fact that language is governed by rules explains why language is learned under the circumstances of actual language acquisition (i.e., rapidly, on the basis of an immature intellect, with ease, without formal instruction but with a uniform result across the speech community). 1.3 Universal Grammar It is obvious that individual languages differ quite radically from each other on the surface, cf. Chinese or Eskimo with English or German. However, closer study immediately reveals that they are surprisingly similar in their deeper properties. Linguists know that behind the diversity of languages and the variability of linguistic structure general regularities can be found. Abstract principles have been discovered that characterize all languages. These are the universals of linguistic theory. All human languages display similar levels of complexity and detail - there is no such thing as a primitive language. One example of a language universal is that the formal operations of all languages are structure dependent. For

5

instance, question formation in English cannot be understood as merely shifting the second word of the sentence in front of the first word as appears to be the case in (1b) vis-à-vis (1a). This gives the wrong result in (2b) vis-à-vis (2a) where the second word is the preposition of. Rather, the operation involved picks out the first (i.e., tensed) auxiliary and moves it in front of the subject noun phrase as shown in (2c). Words group into phrases and it is on the basis of the syntactic phrases that grammatical operations can be defined; they do not make reference to the linear order of words. (1) (2) a. Memories will fade away. b. Will memories fade away? a. Memories of happiness will fade away. b. *Of memories happiness will fade away? c. Will [memories of happiness] fade away?

Another property of language is its creativity - is it always possible to create new complex expressions (words, phrases and sentences) to express whatever the speaker intends to say. Human languages, therefore, are unbounded in scope. 1.4 Language Universals The original idea of language universals was historical in its conception. It stemmed from a school of linguists active in Leipzig in the 1870s to which Brugmann, Leskien, Osthoff, Delbrück, Verner, Paul and Behaghel belonged. They were jokingly referred to as the `Young Grammarians' (Junggrammatiker). The Neogrammarians (to use the more neutral English term) were concerned with what they believed to be a genetic relationship between the individual Indo-European languages and language families. Observable similarities in cognate word forms could be the reflex of a common origin. The method of historical comparison was applied to the phonological and morphological systems of the languages involved in an attempt to reconstruct their earlier stages. The reconstruction was seen as an explanation for the similarities among the words which developed historically from the earlier form via a series of sound laws (`Lautgesetze'). (3) Indo-European *p Sanskrit p pitarpad(no cognate) pasu Latin p pater pes piscis pecu English f father foot fish fee

In this century the notion of linguistic universals has played a major role in linguistic theory as well: Linguistic variance ­ even from a synchronic viewpoint - is thought to be anchored in invariant principles of language structure. Joseph Greenberg, for example, used a restricted number of structural properties to classify a large number of genetically diverse languages into a linguistic typology. One criterion is the order in which the major constituents of a sentence occur, i.e. where a governing constituent (or head) is found in relation to its dependent (or complement). In the following example the head of an English verb phrase or noun phrase precedes its complement; in Korean it follows. (4) English: a. Close the door b. desire for change

6

(5)

Korean:

a. Moonul dadala b. byunhwa-edaehan kalmang

'door close' 'change-for desire'

Universal Grammar (UG) as conceived by Noam Chomsky is based on the notion that language ability is encoded in the human genome and that a common biological endowment restricts the number and type of languages that are accessible to humans. Humans, therefore, have an autonomous language faculty that is task-specific and does not derive from general cognitive abilities. Again, linguistic divergence is postulated to be rooted in invariant principles, but this time of a cognitive (rather than a historical or typological) nature. Universal grammar is a theory of the innate (and highly abstract) principles that determine the general form that a grammar of any language may take on. Knowledge of the principles of UG makes the acquisition of language possible. 1.5 Cognitive Science Along with four other fields (psychology, philosophy, neuroscience, computer science), linguistics makes up the modern field of cognitive science which studies the structure and functioning of human cognition. The basic idea behind cognitive science is that the study of cognition (perception, memory, thought and action) should be a subject of interdisciplinary research, to which the expertise of many traditional disciplines can contribute in a crucial way. Computer science can be used as an analogy that helps us to understand the nature of the human mind. The human mind is considered to be like a mental 'program' and the neurons of the brain make up the 'hardware'. The grammar of a particular language can then be seen as the `software', i.e. a particular program. Knowing how programs and hardware are related in computer science could lead to an understanding of how knowledge and thoughts might be related to the neural structure of the brain. One of the most interesting areas in psychology is the study of how linguistic knowledge is acquired and how it is used in producing and comprehending speech. Linguistic knowledge is a system of information. The relevant questions are: How is information acquired and represented in the mind and how is this information called up and put to use in speaking and understanding? In short, the mind is considered as a system of information processing. Linguistic knowledge is one system that it processes. The contribution of theoretical linguistics to cognitive science is that virtually all human beings learn a language. Language is thus related to human cognition in a fundamental and interesting way. This brings questions like the following to mind: Is the human child preprogrammed to learn a language system? Can other species acquire language? Human language is unique among communication systems in its complexity. Given the natural ability of children for mastering it, it is reasonable to suppose that there is something special about the human brain, either in its capacity or in its structural organization, that makes the use of language possible for humans. The field of neuroscience studies the structure and functions of the human brain that underlie such cognitive abilities as language. 1.6 Animal Communication 1.6.1 Language vs. Communication Language developed in the course of evolution. Can the study of animal communication systems give us an indication of how human language developed? Do animals make use of

7

any system of communication that resembles modern human language? Is language species-specific? Were our pre-homo sapiens ancestors capable of using a primitive language as a precursor to our modern language? These are topics that are still under discussion today. 1.6.2 Bee's Dance When a food source has been discovered, the forager bee flies back to the hive and 'dances' to convey information about it. The dancing bee draws the attention of about a dozen other bees in the dark hive by dancing on the empty cells of the hive which consequently begin to vibrate arousing the other's attention. This causes them to imitate the dance by establishing contact with the dancing bee via their antennas. Apparently imitation is necessary for understanding the message.The bees' dance signals the location of the food source, its quality and its distance from the hive. Distance is conveyed by one of three different dances performed on the wall of the hive: a) the round dance in which the bee circles repeatedly indicates a food source within five meters of the hive, b) the sickle dance in which the bee dances in a curved figure-eight indicates a food source between five and twenty meters from the hive, c) and the tail-wagging dance, finally, indicates a distance further than twenty meters from the hive. In this case the bee wags its tail as it moves forward circling to the right back to the starting point and then repeats the wagging forward motion circling left and back again etc. This is the 'dialect' of the yellow Italian honeybee. The black Austrian honeybees do not have the sickle dance. The round dance doesn't communicate direction (the nectar is close enough to the hive for the bee to be guided by its scent). Bees orient themselves in flight relative to the angle of the sun. When the sickle or tail-wagging dance is performed on the vertical wall of the hive, the top of the hive is understood as the position of the sun in the sky. The angle of both dances relative to the hive's vertical alignment indicates the direction the flight should take toward the food source relative to the sun: The open side of the figure-eight or the angle of the tail-wagging-path signals the direction. The quality of the food source is indicated by the intensity of the dancing. 'Bee language' is innate: Young forager bees on their first flight perform the dances appropriately even though their flight orientation is somewhat imperfect. With more exposure to the sun, the dance is refined within a few hours. An experiment of crossbreeding Austrian with yellow Italian honeybees presents an interesting argument for the innateness of the dance: Bees bearing a physical resemblance to their Italian parent (the yellow marking) performed the sickle dance to indicate intermediate distance 98% of the time. Bees that bore a physical resemblance to their (black) Austrian parent performed the round dance 96% of the time to express intermediate distance and didn't perform the sickle dance at all. The dance pattern chosen appears to be inherited from a certain parent along with other genetic traits. Is the bee dance a rudimentary `language'? Peter Gärdenfors (1996) differentiates signals (whose referents are cued from the immediate context) from symbols (which refer to referents detached from the context of communication). A term used as a signal evokes an action appropriate to the presence of the object it denotes. Symbols do not stand in this direct relationship to their objects, but are abstract conceptions of objects. "To conceive a thing ... is not the same as to react toward it overtly, or to be aware of its presence. ... it is the conceptions, not the things, that symbols directly `mean'." ..."Many animals have intricate systems of signals, for example, the dances of bees. This kind of dance has a kind of `lexicon', it exhibits combinatorial patterns of the elements in the lexicon, and it even

8

satisfies Hockett's `displacement' since the dance refers to a nectar find that is remote from the hive where the dance is performed. However, even if the bees' dances seem to have a kind of grammar, they still consist only of signals. The bees categorize places where nectar can be found in a sophisticated way. The crucial point is that they only use their dances in a cued manner, and thus the dances are not symbols according to my criterion. The same point is made by von Glasersfeld (1976: 222) [who states, S.O.]: `In my terms, the bees do not qualify for symbolicity, because they have never been observed to communicate about distances, directions, food sources, etc., without actually coming from, or going to, a specific location. ... To qualify as a language, the bees' dance would have to be used also without this one-to-one relation to a behavioral response ...' The fact that a language consists of symbols referring to detached representations is a necessary, but far from sufficient, condition to separate language from other forms of communication." 1.6.3 Non-human Primates The larger non-human primates are chimpanzees, bonobos and gorillas. We are closely related to them genetically, in that we share approximately 96% of our genetic material with them. In the 1960s and 1970s several teams of psychologists attempted to teach them language. In 1966 the Gardners began teaching a one year old chimpanzee they named Washoe American Sign Language (ASL). The Hayes' experiment with Viki had shown that the vocal tract of a chimpanzee is not capable of articulating many of the human speech sounds. After six years of training, Viki spoke only four words (mamma, papa, cup and up). Since chimps love to gesture and are very social, the Gardners decided to try a sign language. ASL is a system used naturally by many deaf people and it has a structure that is comparable in many ways to spoken human language. Washoe was with a trainer every day and was continually challenged to use signs. Her first combinations (such as gimme sweet and come open) occured after about 10 months of training. During the next two years she made about 294 different two-sign combinations (such as gimme tickle, more fruit) and even produced four and five sign combinations like you me go out (hurry). Some of the forms used appear to have been inventions by Washoe, like her novel sign for bib and in the combination water bird (referring to a swan), which would seem to indicate that her linguistic system had the potential for productivity. The Nim Chimpsky project carried out by Terrace after the Washoe project raised serious questions about the ability of apes to learn a language. After a long training phase with Nim Chimpsky, Terrace (1985:1012f.) comes to the following conclusion: "By 1980, it was apparent that 'sentences' created by apes could be explained without reference to grammatical competence. My associates and I analyzed approximately 20,000 combinations of two or more signs made by Nim, a young male chimpanzee who, like Washoe, had been reared by his human surrogate parents in an environment in which ASL was the major medium of communication. Superficially, many of Nim's combinations appeared to be generated by simple finite-state grammatical rules (e.g., More + x; transitive verb + me...) Taken by themselves, such combinations provided the strongest available evidence that an ape could create a sentence. Indeed, many of Nim's multi-sign utterances resembled a child's initial multiword utterances ... However, a frame-by-frame analysis of videotapes of Nim's signing revealed that Nim responded mainly to the urgings of his teacher that he sign and that much of what he signed was a full or partial imitation of his teacher's prior utterance(s). ... Although young children clearly imitate many of their parents' utterances, the relative frequency of imitated utterances is substantially

9

lower in children. Further, although the imitative phase in children is transitory, Nim never moved beyond that phase ...." 1.7 References Akmajian, Adrian, R. Demers, A. Farmer & R. Harnish, 1995. Linguistics. An Introduction to Language and Communication. Fourth Edition. Cambridge, Mass.: MIT Press. Fromkin, Victoria & R. Rodman, 1998. An Introduction to Language. Sixth Edition. New York: Harcourt Brace College Publishers. Gärdenfors, Peter, 1996. Cued and detached representation in animal cognition. Behavioral Processes 36, 263-273 O'Grady, William, M. Dobrovolsky & Fr. Katamba, 1996. Contemporary Linguistics. An Introduction. Third Edition. New York: St. Martin's Press. Radford, Andrew, 1997. Syntax. A minimalist introduction. Cambridge University Press. Terrace, H. S., 1985. In the Beginning was the 'name'. American Psychologist, 1011-1028. 2 2.1 (1) Morphology Structure of Words Speakers of a language have the ability to form new words to fit their needs of expression and their conversation partners will understand their novel expressions in the intended meaning. Morphology is the study of this area of grammatical competence, i.e. the internal structure of words and the regularities by which new words are formed. The basic structural units of words are morphemes: Bound morphemes must be joined to other morphemes, rest+less, drive+able, sing+er; they are always parts of words, never words themselves. Free morphemes can constitute words by themselves; they can be simplexes (monomorphemic), as in water, crocodile, or have a complex structure (polymorphemic), as computer, biodegradable, globalization: [N [V [A [N glob(e)] -al] -ize] -ation]. Affixes are bound morphemes that recur in specific structural environments. They can be inflectional (employ+ing, employ+s, employ+ed) or derivational (employ+er, employ+ee, employ+able). Affixes may precede the stem to which they are attached (= prefixes, re+build, un+happy) or follow it (= suffixes, entertain+ment, friend+ship). Infixes are inserted into the stem. Bontoc (spoken in the Philippines), for example, forms a set of complex verbs by infixation: (1) Nouns/Adjectives a. fikas b. kilad c. fusul d. fikas + -um'strong' 'red' 'enemy' > Verbs fumikas kumilad fumusul fumikas. 'to be strong' 'to be red' 'to be an enemy'

(2)

English uses a form of infixation (so-called `expletive infixation' where stems rather than affixes are inserted into a base stem) to form highly emotional expressive forms like abso-bloomin-lutely, in-bloody-credible, Ida-shitty-ho, unfucking-believeable. Circumfixes are discontinuous units that surround a stem (German: ge+schaff+t, ge+worf+en / Ge+schnarch+e, Ge+tanz+e, Ge+jammer+e, Ge+frag+e). A morpheme to which an affix attaches is a base.

10

(3)

Most base morphemes are free, but some can be bound (so-called 'cranberry' morphemes). These are not meaningful in isolation but are meaningful only in combination with another specific morpheme, cf. cran+berry vs. straw+berry, blue+berry. Other examples: lukewarm, grateful, inept, debunk; German: Unrat, Schornstein, niedlich, hämisch. English also displays bound stems: perceive, receive, deceive; remit, permit, submit, admit, transmit. Germ.: Fanatiker, Fanatismus, fanatisch. Lexicography, the making of dictionaries, registers certain facts about wellaccepted words of a language. Morphology, a sub-discipline of linguistic theory, attempts to model a speaker's knowledge of the words he knows and ­ based on this knowledge of the vocabulary of his language - his ability to form new words. a. Abbreviation (Acronyms): AIDS, UCLA, CNN, PC, PR, www, CEO (chief executive officer), CD-ROM (compact disc-read only memory), MB (megabyte), ISDN (integrated services digital network), SETI (Search for Extraterrestrial Intelligence), ATM (automated-teller machine), SMS (short messaage system); German: ZP, EDV, Bafög, DGfS. b. Shortening (clipping) is a process whereby a new word is created by shortening a polysyllabic word: fax (facsimile), condo, porno, promo, techno, sitcom, veggie. Older: prof, exam, ad, bike, gas, phone. German: Abi, Uni, Abo, Demo, Euro, Stip. c. Contamination (blending): stalkarazzi (stalk+paparazzi) guesstimate (guess+estimate) wegotism (we+egotism) netiquette (Net+etiquette) gundamentalist (gun+fundamentalist) slanguage (slang+language) informavores (information+carnivore) "although he delivers Bond-mots with requisite panache, Brosnan tends to play the part straighter" the Coca-Colonization of global culture German: Ostalgie, Formularifari, Bonnflikt, Kurlaub, Berlinienbus. d. In `neo-classical compounding' the units of the combination are not native stems but non-native roots (mostly from the classical languages Latin and Greek) such as bio-, auto-, tele-, -scope, -ology, -phile etc. Such roots must be considered bound elements in English, but are at times difficult to distinguish from prefixes and suffixes, cf. automobile vs. autosuggestion, autotoxin and lexicology, semiology, philology, ethnology etc. German: Tonsillitis > Telefonitis, Schriftstelleritis, etc. e. Coining entails inventing a new arbitrary pairing of sound and meaning: googol '10100', scag 'to shot heroin'. (Extension: to spam `bombard an enemy's computer with thousands of junk messages': [spam = `spiced pork and ham', i.e. a type of `meat distributed in cans' > spam(mail) = `send phenomenal amounts of mail'.) f. Generification uses a brand to name the product: kleenex, xerox, Mac; Tempo.

(4)

(5)

11

g. Reim-/Ablautdoppelung: zigzag, knick-knack, riffraff, tip-top, crisscross; German: Schickimicki, Techtelmechtel. (6) The two most productive types of word formation in languages like English and German are compounding and derivation. These are regular concatenative processes central to the language system. a. In compounding, individual words are joined together to form a compound: fax machine, information highway, data traffic, punk rocker, silicon realm. b. Derivation is the process by which a new word is formed from a base, usually through the addition of an affix. i. Suffixation: fix+able, rip+en, soul+less. a) read+able, think+able, wash+able, mention+able, understand+able. b) [A V + able] ii. Prefixation: ex+spouse, un+clear, re+think, anti+beef. a) un+able, un+aware, un+fair, un+fit, un+kind, un+stable, un+steady [A un + A] = negation. b) un+do, un+glue, un+lock, un+learn, un+pack, un+thread, un+ravel [V un + V] = reversal of action. An apparently non-concatenative process of word formation that is very productive in English and German is conversion (or zero derivation). It assigns an already exisiting word to a new syntactic category. a. V to N: to call > a call, scream, drive, purchase, walk. b. N to V: a boss > to boss, to shovel, to pile, to skyrocket, to handcuff c. A to V: yellow > to yellow, narrow, dim, mature, sour. Productive examples: "called on the Gore campaign not to lawyer the race to death". Reduplication, productive in many languages, is a process of duplicating all or part of the stem. a. In Tagalog a copy of the first consonant + vowel sequence of the root is added initially to the stem to form the future tense: (2) a. tatakbuh 'will run' < b. lalakad 'will walk' < takbuh 'run' lakad 'walk'

(7)

(8)

b. In Indonesian the entire word is repeated to form a collective: anak anak 'all sorts of children'< anak 'child'; maõga maõga 'all sorts of mangoes'< maõga 'mango'. c. Samoan: (3) manao matua' malosi punou atamaki savali 'he wishes' 'he is old' 'he is strong' 'he bends' 'he is wise' 'he travels' mananao matutua malolosi punonou atamamaki savavali 'they wish' 'they are old' 'they are strong' 'they bend' 'they are wise' `they travel'

d. Reduplication is limited in English and German: goodie-goodie; Eff-Eff, Tamtam, Pinkepinke.

12

(9)

Backformation attests to the psychological reality of word formation rules. This is a phenomenon in which a productive rule of word formation is reversed, i.e. an apparently complex word is `deaffixized.' Historically the complex words pedlar, beggar, scavenger, swindler, editor, burglar and sculptor all existed in the language before the verbs to peddle, beg, scavenge, swindle, edit, burgle and sculpt. Speakers assumed the ­er/-ar/-or at the end of the noun was the agentive suffix -er and substracted it from the putative complex word to arrive at its 'base'. The pattern of agentive nouns derived from a verbal base is extremely productive in English and, hence, salient to the speaker: to walk > walk+er, to cheat > cheat+er, to think > think+er, to send > send+er etc. Backformation occurs in such cases because the word formation regularities are in a sense 'misapplied' by speakers. More accurately, they analyze a monomorphemic word as if it were a complex word on the basis of the regular patterns of the language and use this analysis to arrive at a new stem. But speakers can also extract an affix from an affixed simple base form that has undergone compounding to arrive at a new complex base, cf: to backform < backform+ation, to self-destruct < self-destruct+ion, to aircondition < aircondition+er. Action nouns formed from verbs by the suffix ­ation (form > formation) have combined with a first constituent yielding a nominal compound that is transformed into a new complex verbal stem. Cf. the recent introspect from introspect+ion as in `ask native speakers to introspect about the nature of the process.' An older example is the verb televise which results from misanalysis of television (tele+vision) as televis+ion > to televise.

(10)

Analogy is a force in word formation where one morpheme is subtracted from a complex base and replaced by another that is more appropriate to the situation. a. e-card < e-mail, drinkathon < marathon, slow food < fast food, cheat wave < heat wave. Patterns based on one creation can arise: b. space+scape, moon+scape, city+scape, dream+scape, street+scape < land+scape; c. First Marriage, First Couple, First Daughter, First Dog, First Golfer, First Flier, First Niece, First Twins < First Lady; German: Vorritt, raumkrank, Henkerszigarette, Verleser. Games a. fender bender, shock locks, chick flick, poop scoop, boob tube. b. Got it Maid [Maid in Manhattan] Cruising to the top [Eyes Wide Shut] Claus encounter [Kelly Osbourne as Santa Claus] c. Many see the happiness industry as a case of the bland leading the bland. It was a case of the blonde leading the blonde in Milan [Donatella Versace and Britney Spears] Motivation and compositionality of meaning. Complex words are compositional when their meaning is explainable as the product of the meaning of their constituent parts. Once coined, however, many words become fixed units of the lexicon and tend toward some degree of idiosyncrasy. Cf. the A+N compounds old clothes, big city, blind spot, black top, green onion, fast lane, hot line, happy hour.

(11)

(12)

13

(13)

Semantic drift occurs when a word takes on some additional feature of meaning independent of its morphological origin and in addition to the compositional meaning of its parts. (4) a. b. c. d. e. divorce law, tennis match, liquor store pigtail, skyscraper, rubber check paper clip, eggplant, drugstore cupboard (|kYbcrd|), lord (OE: hlaf 'loaf' + weard 'warden') cranberry motivated complex metaphorical construction not entirely motivated demotivated unique morpheme

2.2 (14)

Compounding A compound is formed by combining two stems. The second stem functions as the head which determines the semantic and morpho-syntactic properties of the whole word, while the first stem somehow modifies the head. Between the two, a linking morpheme can appear. These features are most clearly illustrated in German: die Dorfstraße where Straße (fem.) is the head and das Straßendorf with Dorf (neut.) as the head and -n- as a linking morpheme without a clear function in the compound. (5) N Dorf N N Straße N Straße(n) N N Dorf

In English the compound hot line (A+N) is a noun because line belongs to the category noun. A compound can be part of another compound: (6) A hot N N line Yuppie N N phone N N N sex

The individual constituent groups guide the interpretation of the compound: (7) a. b. c. d. [[breast cancer] gene] [Friendster [bulletin board]] [[safe sex] [advertising campaign]] [[short message] system [ticket sales] service]

The structure of compounds is a simple binary concatenation. The primary challenge for the theory of grammar is how the meaning of novel (non-lexicalized) compounds is to be explained.

14

(15)

The traditional classification of compounds was based on 3 general categories taken over from Sanskrit grammar: a. Endocentric or determinative compounds are combinations of stems in which the compound denotes a subset of the denotation class of the head (A + B = B). The first constituent serves to restrict the meaning of the second. This type is the most regular and productive. Ex.: beer bottle 'a bottle for beer', computer program, street sign. German: Waldweg, Straßenschild. b. Exocentric or possessive (bahuvrihi) compounds are similar to endocentric compounds in that the first constituent restricts the meaning of the head, however the whole compound denotes a third entity which typically displays the property denoted (A + B = C with AB). Ex.: greybeard 'someone/thing (e.g. a seal) with a grey beard', skinhead, paperback. German: Milchgesicht, Krauskopf. c. Copulative (or dvandva) compounds contain two constituents of the same category that denote attributes applying equally to a third entity (= C has A + B). If the compound is novel, it is usually possible to change the order from A+B to B+A. Ex.: kosher-cajun/cajun-kosher (A+A) and scholar-activist/activist-scholar (N+N), cowboy-vet, schoolteacher-lawyer. German: Opfer-Zeuge. Composition is a productive process creating novel words whose meaning is generally open to more than one interpretation, cf. actress daughter `(my) daughter & actress' (= copulative) or 'daughter of an actress' (= determinative); sun spot `spot on the sun/caused by the sun/in the shape of the sun'; lawyer joke `joke about/told by lawyers', etc (= all determinative). Marchand (1969:18) states: All compounds are explainable on the basis of syntactic relations underlying them in sentences. He postulates that compounds arise on the basis of an underlying sentence by means of a topicalization rule that picks out the topic and assigns it the status of determinatum (or: head) of the compound. Further reduction of the sentence isolates the determinant (i.e., nonhead) and allows it to `expand' the determinatum into a compound syntagma. Both members of the compound retain their original grammatical relations in the reduced construction. (8) a. Someone eats an apple. b. apple eater c. apple eating d. eating apple topic: subject predicate object

(16)

(17)

Marchand (1969:22) differentiates two principal types of compounds, verbal nexus and non-verbal nexus or primary compounds: Verbal nexus combinations are those which are made up of a verbal and a nominal element provided the two form a direct syntagma in an underlying sentence. Primary compounds are ... combinations in which a verbal nexus is not overtly expressed but implicitly contained ... (9) a. car dealer b. car thief `one who deals in cars' `one who steals cars'

The problem is to determine the relevant relation that holds between the head of the primary compound and its first constituent: A + relation + B; what is the appropriate 'relation'?

15

(18)

Fanselow (1981) is also of the opinion that compounds are interpreted on the basis of an explicit 'verbal' or an implicit 'non-verbal' (and, therefore, `inferred') relation. a. Verbal compounds display a specific and grammatically characterizable range of semantic interpretations. Ex.: sheep shearer, lion tamer, lawn mower, ink blotter. b. Non-verbal compounds allow an indefinite number of relations between the head and non-head constituents and consequently can't be characterized by the grammar alone (in the proper sense of 'grammar'). Their interpretation draws on factors that go beyond linguistic knowledge and are part of our general cognitive knowledge of the world or are anchored in the situation of use. Ex.: space scientist, sun spot, banana boat, leather tool. The interpretation of verbal compounds is based on the argument structure of the head. a. In a compound like sheep shearer, the head shearer is a complex noun based on the verb shear which is a two-place predicate ('someone SHEARs something'). The initial constituent can be interpreted as object of this predicate 'someone who(/something that) shears sheep'. Also: lion tamer, ink blotter; German: Taxifahrer, Taubenberinger, Seitentrenner. b. This applies to relational nouns as well: street corner where street completes the denotation of corner 'corner of a street'; also table top, town mayor, shirt sleeve, cigarette butt. German: Bahnsteigkante, Professorensohn. c. That this type of interpretation is based on the grammatical properties of the head constituent can be shown by the fact that certain forms are excluded when the first constituent doesn't meet the semantic restrictions of the head. Ex.: *stone drinker, *beauty shearer, *peace collector, *bravery-opener, *tree corner. d. This interpretation competes under certain circumstances with non-verbal interpretation in which case the first constituent takes on the function of a modifier: news reporter `one who reports news' vs. newspaper reporter 'one who reports for a newspaper', book publisher vs. Web publisher. Interpretation of non-verbal compounds (after Fanselow (1981)): a. A compound made up of two common nouns A and B may only entail relations that derive from the meaning of these two constituents (in a broad sense). (1) Supply a basic relation that exists independently of the meaning of the constituents such as AND, PART_OF, PLACE_OF. (i) (Russia's) gangster-businessmen, cowboy-vet, kidnapper-killer, exploreranthropologist, scoundrel savior, actor-environmentalist, philosopher-novelist, poet-victim, nerd genius. (ii) broom stick, mountain peak, arrow head, door knob, piano keys, tire rim, shirt sleeve, seatback. (iii) space junk, cave cemetery, ground transportation, shoreline, prairie dog. (2) Induce a prominent relation from the initial or final constituent and apply it to the meaning of the other constituent. Ex.: hymn-book ('contain'), job tension ('caused by'), meat knife ('cut'), oil company (`deals'), lung cancer (`affects'). (3) Modify the induced relation locally. Ex.: lowland fox ('live') lobster pot ('cook'), neck-tie ('wear'), air traffic ('move'), football stadium ('play'), flower bed ('grow'), star wars (`fight'). (4) Induce from the first constituent a property that can be applied to the second constituent. Ex.: sponge cake (texture), sperm whale (shape), snail mail (speed), snake dance (figure), skunk cabbage (smell), razor clam (feel), grandfather clause

(19)

(20)

16

(age), graveyard shift (`dead'). b. The point is that a relevant relation between A and B is exploited in interpreting such structures (REL (A,B)). The actual content of this unspecified relation isn't dependent on grammatical factors and consequently does not enter into the grammatical characterization of the compound but is supplied by extragrammatical knowledge (from the context, by our associations etc). c. In this system every such compound is potentially ambiguous. Ex.: senator actor: REL(senator, actor). 1) REL = AND 2) REL = 'portrays' ... 3) REL = 'established like a senator', 'older like a senator', 'experienced like a senator', 'appearance of a senator', and so on. In Fanselow's system, every compound is potentially ambiguous but the ambiguity is meaning-driven and contained within the limits of plausible connections between the constituent meanings. For example, a space scientist could be a scientist who `studies' space (rule 2) or a scientist `working in' space (rule 3) and a hedgehog gene could be an individual unit of the genome of a hedgehog (rule 2) or a gene that has the `shape' of a hedgehog (rule 4). Unwanted meanings, Fanselow argues, don't arise. If one were to interpret German Haushund by means of rule 1 using the AND-relation, a contradiction would occur since there is no possible referent of the compound that can be both a house and a dog simultaneously. Such a contradiction of our world knowledge disqualifies this meaning as a possible interpretation. Furthermore, Küchenmesser or Engl. kitchen knife can only be interpreted locally; interpretation by means of the stereotype of a knife (i.e. that it `cuts') yields an absurd result. On the other hand, Bergdenkmal and factory violin could - in the right context - be interpreted as a `monument that depicts a mountain' or a `violin for playing in a factory'. The rule system developed allows these contextual readings as well; however, Fanselow considers them contextually dependent as opposed to the contextually independent readings generated in isolation of a context. It is important to note that for all types of interpretation the structure remains constant: A very simple structure allows for the many different types of interpretations that arise. (10) N logician broom beach pain 2.3 (21) N N philosopher stick towel reliever

Derivation by Suffixation Some productive suffixation patterns make crucial reference to the argument structure of the base verb (as in the verbal compounds). The verb enforce denotes a situation in which someone enforces something: He enforced the law. Formally: [e

17

INST [x ENFORCE y]], to be read 'an event e instantiates the proposition x enforce y'. The verb has an argument structure: y x e [e INST [x ENFORCE y]]. a. The suffix -ment attaches to the verbal base forming a noun that denotes the event of enforcing. The derived noun binds the e (= event) position in the above formula: enforcement: y x [e INST [x ENFORCE y]]. Also: appointment, harassment, development, enactment, appeasement. b. -er attaches to the base forming a noun that refers to the subject of the verb. It picks out the x position in the formula: enforcer: y [e INST [x ENFORCE y]]. Also: gambler, dealer, supporter, designer, listener. c. -able forms adjectives referring to the original object (= y) of the verb as their subject: The law was enforceable. Enforceable: [e INST [x ENFORCE y]]. Also: exchangeable, determinable, measurable, understandable, attainable. (22) Other suffixation patterns are based on a very general relation that carries quite a bit of implicit information (as is the case with non-verbal compounds). The suffix ­ y, for example, combines with a noun to form an adjective with the meaning `characterized by N', cf. (11a). In (11b) further nuances have crept into the meaning. (11) a. b. dirt+y, dust+y, mess+y, thorn+y, leaf+y, cloud+y, sun(n)+y, grass+y, water+y, sugar+y, grain+y taste+y `tastes good', word+y `excessof words', price+y `high in price'

The suffx ­ish also combines with nouns to forms adjectives that carry approximately the same meaning of `belonging to' the content of the base noun (as also in English, Irish, Swedish, etc.). At times it can take on a negative connotation (perhaps partially due to the content of the base noun as in foolish) and it can also add on to complex bases. (12) a.child+ish, boy+ish, girl+ish, fever+ish, styl(e)+ish, imp+ish, knav(e)+ish b. fool+ish, snob(b)+ish, devil+ish, goul+ish, prud(e)+ish c. sheep+ish `timid', book+ish `fond of books' d. at-home+ish, out-door+ish, public-school+ish, schoolboy+ish, standoff+ish, old-maid+ish

The suffix ­esque also forms adjectives from nouns with the idea of being characterized by, but the pattern is more restricted in meaning to `having the style of'. (13) pictur(e)+esque, carnival+esque, Rembrandt+esque, Kipling+esque, Kennedy+esque, Great Gatsby-esque Garbo+esque,

Some suffixal patterns must be considered active even though they are no longer completely productive such as the previous patterns are. One example is ­age whose forms documented in (14) carry the meaning `collection of' or `duty, fee or place connected with the collection' when combined with a noun and `result of' when suffixed onto a verbal base: (14) a. parsonage, brokerage, orphanage b. baggage, mileage, tonnage, voltage, yardage c. anchorage, postage, poundage, storage, stowage

18

d. breakage, coverage, spillage, wreckage, sewage, drainage 2.4 References Bauer, Laurie, 1983. English Word-Formation. Cambridge: Cambridge University Press. Fanselow, Gisbert, 1981. Neues von der Kompositafront oder zu drei Paradigmata in der Kompositagrammatik. Studium Linguistik 11, 43-57. Marchand, Hans, 1969. The Categories and Types of Present-Day English Word Formation. München: Beck Olsen, Susan, 1986. Wortbildung im Deutschen. Stuttgart: Kröner Spencer, Andrew, 1991. Morphology Theory. Oxford: Blackwell. Syntax Constituent Structure Syntax deals with the structure of sentences. Not only sentences, but any complex thing has structure which basically means 1) that it is made up of different parts (or: constituents), 2) that these parts are of different kinds (or: categories), 3) that the categorized parts are put together in a specific way, and 4) that each constituent has a distinct role in the function of the complex structure. The individual constituents can themselves be complex in that they may contain other constituents. We then say that the structure is hierarchical. A bicycle is not simply a random collection of individual pieces (tires, spokes, chain, seat, handlebars etc.) as we might find in a sculpture, for instance, but the individual parts are ordered together in special ways that allows them to perform a specific function with regard to the function of the whole object. A sentence is a structured linguistic object. Words are grouped into successively larger phrases or structural units called constituents. The constituent structure of a sentence can be made apparent by a tree diagram (cf. Radford 1981): (1) This boy must seem incredibly stupid to that girl.

3 3.1 (1)

(2)

This

boy

must

seem incredibly

stupid

to

that

girl

(2) *Incredibly girl that seem to must boy stupid this. * this (3) boy must seem incredibly stupid to that girl

Similarities and differences between constituents are described in terms of categories. Boy, girl belong to the category noun; the, a, this are determiners; seem, run, believe are lexical verbs; must, can are modal (auxiliary) verbs; to, with, in are

19

prepositions; incredibly, really are adverbs; stupid, mad, rich are adjectives. (4) Syntactic phrases are built around a key word or head. [this boy] is a noun phrase (NP) because its head is a noun. [incredibly stupid] is an adjective phrase (AP) because its head is an adjective. [to that girl] is a preposition phrase (PP); its head is a preposition. [seem incredibly stupid to that girl] is a verb phrase (VP) whose main constituent is a verb and the whole sequence in (2) is a clause or sentence (S). The tree below is called a phrase marker because it marks the hierarchical grouping of words into phrases.

(3)

This boy must seem incredibly stupid to that girl. S NP M N V Adv VP AP A P PP NP

D

D This boy must seem incredibly stupid to

N

that girl

S = clause, NP = noun phrase, M = modal, VP = verb phrase, D = determiner, N = noun, V = verb, AP = adjective phrase, PP = prepositional phrase, ADV = adverb, A = adjective, P = preposition (5) Sentence structure is hierarchical: It is made up of phrases of various different categorial types that contain phrases of different categories. Structural tests yield evidence for hierarchical constituent structure as seen in the above example (3). Syntactic movement: a. Certain phrases can be preposed in English for emphasis: 1. a. I can't stand [NP your elder sister]. b. [NP Your elder sister] I can't stand. 2. a. They said it would be [AP very exciting], b. and [AP very exciting] it was. 3. a. He is going to be leaving for Paris [ADVP very shortly]. b. [ADVP Very shortly], he is going to be leaving for Paris. 4. a. John ran [PP up the hill]. b. [PP Up the hill] John ran. 5. a. I won't [VP give in to pressure]! b. [VP Give in to pressure], I won't! Only a whole phrase can be preposed: 1'. *[NP Your elder ___] I can't stand ___ sister.

(6)

20

Only phrasal constituents may be preposed: 6. John [V rang up] [NP his mother]. 6'. *Up [NP his mother], John [V rang __]. (In this case up belongs to the verb (= to ring up); it isn't the head of a prepositional phrase *up his mother.) b. The same regularities are involved in NP postposing: 7. He explained [NP all the terrible problems he had encountered] to her. 7'. He explained ___ to her [NP all the terrible problems he had encountered]. 7".*He explained [NP all the ___] to her [___ terrible problems he had encountered]. Only phrasal constituents (i.e. whole phrases) can undergo syntactic movement. (7) Pronominalization: a. Pro-forms refer back to antecedents that are constituents. 1. You must hate [NP this long boring lecture on constituent structure]. 2. You must hate [NP it]. 3.*You must hate [NP it on constituent structure]. 4.*You must hate [NP this long boring it on constituent structure]. b. Interrogative pronouns. 1. [S [NP The president] [VP went [PP into the meeting] [PP on roller blades]]]] 2. Who went into the meeting on roller blades? 3. Where did the president go on roller blades? 4. How did the president go into the meeting? 5. What did the president do? Coordination: 1. He owns [NP a cat] and [NP a dog]. 2. I met your [N mother] and [N father] yesterday. 3. Is the cat [PP under the couch] or [PP in the hallway]? 4. He speaks [ADVP rather slowly] and [ADVP somewhat awkwardly]. 5. She walked [P up] and [P down] the road. 6. He may [VP stay home] or [VP go out to a movie]. 7. *John rang [? up his mother] and [? up his sister]. Only constituents can be conjoined. 8. John wrote [PP to Mary] and [PP to Fred]. 9. John wrote [NP a letter] and [NP a postcard]. 10.*John wrote [NP a letter] and [PP to Fred]. Only constituents of identical categories can be conjoined. Adverbial placement: Adverbs like certainly are sentence adverbs (they can only be attached to an S node); adverbs like completely are VP-adverbs (they are only attachable to a VP-node). (4) a. Certainly/*completely, the team can rely on my support. b. *The certainly/completely team can rely on my support. c. The team certainly/*completely can rely on my support. d. The team can certainly/completely rely on my support. e. *The team can rely on certainly/completely my support. f. The team can rely on my support certainly. g. The team can rely on my support completely. h. The team can rely on my support completely certainly.

(8)

(9)

21

i. *The team can rely on my support certainly completely. S NP M V P VP PP NP

the team can rely on my support Possible positions for S-adverbs like certainly are: 1. [S * [The team] * [can] * [VP rely on my support] * S]

Possible positions for VP-adverbs like completely are: 2. (5) [S [The team] [can] [VP * rely * on my support *] S] S-adverbs only attach to an S node; VP-adverbs only attach to a VP-node.

1. (Certainly) the team (certainly) can (certainly) rely on my support (certainly). S NP M V P (c.) the team (c.) can (c.) rely on VP PP NP my support (c.)

2. The team can (completely) rely on my support (completely). S NP M V P the team can (c.) rely (c.) on VP PP NP my support (c.)

3. The team can rely on my support completely certainly.

22

S NP M V P the team can VP PP NP completely certainly

rely on my support

4. *The team can rely on my support certainly completely. S NP M V P the team can rely on VP PP NP certainly completely *

my support

The crossing of branches is not permitted. (10) Levels of categories. There are word-level categories (N, V, P, A, ADV etc.) and phrase-level categories (NP, VP, PP, AP, ADVP). A word-level category always establishes a syntactic phrase. 1. Cars can be useful. 2.* N Cars 3. NP N S M can S M V VP AP A Cars can be useful V be A useful

- Cars has the distribution of other NPs: Can cars be useful? I enjoy cars. - Cars can be expanded to encompass further categories: Fast cars can be useful. - A simple noun can prepose like a NP: Cars I don't like. - A simple noun coordinates with a NP: I like cars and other expensive vehicles.

23

(11)

Ellipsis argues for the existence of the (rather abstract) category VP: The material omitted in the following sentence is identical to a preceding VP: (1) (2) a. John won't help me with the dishes, but his brother will help me with the dishes. a. *John won't put the vodka into the drink, but his brother will put the vodka into the drink. b. *John won't put the vodka into the drink, but his brother will put the vodka into the drink. c.*John won't put the vodka into the drink, but his brother will put the vodka into the drink. d. John won't put the vodka into the drink, but his brother will put the vodka into the drink. S NP D N M V D VP NP N P PP NP D his brother will put the vodka into N

(3)

the drink

In a structure in which a VP is preposed, the VP- adverb moves with the VP but the S-adverb remains behind. Also, when a VP is elided as in the example above, the VP-adverb is deleted along with the VP but the S-adverb remains: (4) It is strange, that the team can rely on his support completely, but ... a. [VP rely on his support completely], they certainly can. b. they certainly can [VP rely on his support completely].

(12)

Lexical vs. structural ambiguity. 1.a. The ball is in the next room. b. John jumped on the trunk. c. Ich fand das alte Schloß nicht. d. Der Hahn läuft. 2.a. old men and women b. John dislikes boring relatives. c. The chickens are ready to eat. d. Linda looks very hard. 3.a. [old [men & women]] b. [[old men] & [women]] 4.a. John dislikes [NP [AP boring] relatives] b. John dislikes [VP [V boring] [NP relatives]] 5.a. The chickens are ready [S (chickens) to eat (s.t.)] b. The chickens are ready [S (s.o.) to eat (chickens)] 6.a. Linda [VP [V looks] [AP very hard]]

24

b. Linda [VP [V looks] [AdvP very hard]] 3.2 (13) Complementation A sentence is divisible into constituents that belong to different categories. Each constituent has a specific function in the sentence. (1) a. Subject (NP): b. Predicate (VP): c. Direct Object (NP): d. Indirect Obj. (NP): e. Obj. of Prep. (NP): f. Predicative (AP/NP/PP): g. N modifier (A): h. A modifier (Adv): i. VP modifier (Adv): j. VP modifier (NP): k. VP modifier (PP): The clown refused to perform. The clown refused to perform. No one likes soggy chips. I sent my sister an e-mail. I sent an e-mail to my sister. Jack and Jill ran up the hill. They dealt with the crisis. Sue is shy / a pianist / at the office. his [funny jokes] his [rather funny] jokes Sue speaks quickly. John went to the disco yesterday. This species of lizard lives in water.

It is important to distinguish between category and function. N, NP, V, VP, P, PP etc. are (lexical and syntactic) categories. Subject, object, predicative, modifier etc. are functions that a category takes on in a specific sentence. (14) Verbs fall into various subcategories depending on whether they require a complement (a phrase to complete their meaning) inside their VP and, if so, what type of complement they require. The verb is said to govern its complement: (2) a. Intransitive: b. (Mono)transitive: c. Ditransitive: d. Copular: Sue sneezed/arrived/vanished/died. My dog has fleas./John drank a beer. He sent the men a check./I handed the tourist a map. She became a lawyer./Fred is rather odd./Martha stayed at the hospital. e. Complex transitive: Sam dyed his hair green./They elected the man president./She kept her car in the garage. f. Prepositional: We rely on your support./We coped with the problem. g. Sentential: We thought that John knew the answer.

Subcategorization frames specify the environment of the verb: (3) a'. sneeze: [VP ___] b'. drink: [VP ___ NP] c'. send: [VP ___ NP NP] d'. dye: [VP ___ NP AP], elect: [VP ___ NP NP], keep: [VP ___ NP PP] e'. is: [VP ___ NP/AP/PP] f'. rely: [VP ___ PPon] g'. think: [VP ___ S]

(15)

The following sentences are ill-formed. However, 1. and 2. are ungrammatical because they violate the subcategorization of the verb, whereas 3. and 4. are

25

grammatical but nonsensical (in normal contexts) because they violate the semantic restrictions placed by the verb on its complements (its selectional restrictions). (4) a. *They described. b. *Sue stomped the flowers. c. §The boy frightened sincerity. frighten: [+/-abstract ___ +animate] d. §Honesty kicked the dog. kick: [+animate ___ -abstract]

(16)

A distinction must be made between complements and modifiers. All verb phrases can optionally include modifiers. A modifier is a phrase that is not subcategorized by the verb and hence the verb doesn't exert selectional restrictions on it. It is an optional element that gives additional (i.e. not essential) information about the activity denoted by the verb and hence can occur with most verbs. In the sentences below the PP with in is not a complement of the verb; the sentences are complete without it. It functions as a local modifier: (5) a. Sam sneezed in the hay stack. b. Max spotted wildcats in the field. c. Ed was extravagant in the store. d. William gave Sue a note in the hall. e. Bill made his mother mad in the theater. f. Stan talked about Martha in the meeting. intransitive transitive copular ditransitive complex transitive prepositional

Modifiers express a wide range of adverbial notions such as manner, means, purpose, reason, place and time. The sentence in (8) has 5 different modifiers. (6) (7) (8) Max buttered the toast in the bathroom with a knife at midnight. Out of disgust, Max turned off his computer in order to remain sane. He guzzled beer noisily at the counter in the bar until midnight every Saturday.

An adverbial modifier denotes a function that can be expressed by such different categories as PP, AdvP, NP or S. A VP adverbial modifies the entire VP (verb plus complements). It is adjoined to VP under the larger VP. Two levels of VP distinguish between complements as sisters of the verb and modifiers as sisters of the VP. Complements are sisters of the verb group. Adverbial modifiers are sisters of the verb phrase. (9) VP Vgroup (17) VP < modifier>

Tests for complements. The expression do so serves as a proform that replaces material inside the verb phrase (the verb and its complements): 1. Sue repaired the car in the garage. 2. Sue put the car in the garage.

26

3. NP

S VP1 VP2 V Sue repaired S NP V NP P Sue put the car in VP PP NP the garage NP the car P in PP NP the garage

4.

5. 6. 7.

a. Sue repaired the car in the garage and Ed did so too. b. Sue repaired the car in the garage and Ed did so in the driveway. a. Sue put the car in the garage and Ed did so too. b. *Sue put the car in the garage and Ed did so in the shed. S NP VP1 V Sue Ed NP in the garage in the driveway VP2 PP

repaired the car [VP1 --- did so (too) ---- ] S

8. NP

VP1 V NP the car did so too PP in the garage ------------ ]

Sue Ed

put [VP1 -------

27

9. Sue repaired the car in the garage on Monday and Ed did so too. 10. NP VP2 VP1 V Sue Ed NP on Mon. -------- ] PP S VP3 PP

repaired the car i.t.garage [VP -------------did so (too)

In addition to VP adverbials, there are also sentence (S) adverbials. 11. a. Fred admitted everything frankly. b. Fred admitted everything, frankly. c. Frankly, Fred admitted everything. a. Max can only do the tango rather awkwardly. b. Max can only do the tango, rather awkwardly.

12. (18)

Optional complements and ellipsis of complements in specific contexts. Read is a monotransitive verb used intransitively; give is a ditransitive verb used monotransitively: 1. 2. Max was reading (a book). [Conversation about Max's birthday presents:] Sue gave (Max) the computer.

3.3 (19)

The Verb Complex The verb group: Every non-elliptical verb group contains a lexical verb as its head which can be modified by auxiliary verbs, cf. He could have guessed the solution. There are two kinds of auxiliaries: - the primary verbs (be, have, do) - and modal verbs (can, could, may, might, shall, should, will, would, must). The primary verbs can also function as lexical verbs. a. In questions the first (tensed) auxiliary can move in front of the subject NP; the lexical verb in VP cannot. If there is no auxiliary verb present, the 'dummy' (i.e. lexically empty) do is inserted: 1.a He can dance. 2.a He is dancing. 3.a He has danced. 4.a He danced. 1.b Can he dance? 2.b Is he dancing? 3.b Has he danced? 4.b *Danced he?

4.c Did he dance?

28

b. In negative sentences, the negative particle not is attached to the first (tensed) auxiliary. Again, if there is no auxiliary verb present, the `dummy' do is inserted: 5.a He can dance. 6.a He is dancing. 7.a He has danced. 8.a He danced. 5.b He cannot dance. 6.b He isn't dancing. 7.b He hasn't danced. 8.b *He danced not.

8.c He didn't dance.

c. The first (tensed) auxiliary can be emphasized to indicate the truth of the proposition or state of affairs expressed in the sentence. 9.a He can dance. 10.a He is dancing. 11.a He has danced. 12.a He danced. 12.c He DID dance. 13.a He can dance. 14.a He is dancing. 15.a He has danced. 16.a He danced. 9.b He CAN dance. 10.b He IS dancing. 11.b He HAS danced. 12.b *He DANCED. 13.b Can't he? 14.b Isn't he? 15.b Hasn't he? 16.b *Danced he not?

16.c Didn't he?

d. The structure of the auxiliary complex consists of the following constituents: 1. TENSE + MODAL + PERFECT + PROGRESSIVE

These categories are optional, any combination is possible, but they always appear in the order given. Each constituent can only appear once. e. Primary verbs have counterparts that are lexical verbs. Need, for example, can act as an auxiliary or as a lexical verb: 2.a He needs to go. 2.b Does he need to go? 2.c He need not go. 2.d Need he go? 2.e He needs a drink. 2.f. Does he need a drink? 2.g *Need he a drink? 2.h *He need not a drink. (20) a. Modal verbs are distinguished from primary and lexical verbs in that they don't show subject-verb agreement or have non-finite forms. - Lexical verbs: gives, gave, to give, giving, given, - A verb following the modal occurs in its basic stem form: 1. He will dance. = modal + bare stem

b. The perfect auxiliary verb can carry tense but appears untensed when it follows a modal. It requires a perfect participle following it. 2. He has danced. = perfect + perfpart

29

He will have danced.

= modal + bare stem + perfect + perfpart

c. Progressive aspect is expressed with a form of be and the progressive participle: 3. He is dancing. = prog + prespart He will be dancing. = modal + bare stem + prog + prespart He has been dancing. = perf + perfpart + prog + prespart He will have been dancing.= modal + bare stem + perf + perfpart + prog + prespart (21) Is the italicized form of the verb be in the following sentences an instance of the lexical verb (copula) or the progressive auxiliary? 1. His behavior may be odd. 2. It is becoming noticeable. 3. John is being nice. 4. He could have been the right man. 5. Sue has been working. 6. The answer is surprising. 7. Bill was drunk by midnight. (22) AUX Tense Modal Perf past past pres pres pres past 3.4 might has is may could have have been Prog V boiled know hidden learning finished telling VP NP the broccoli the answer his keys French the project the truth

References Burton-Roberts, Noel, 1997. Analysing Sentences. London: Longman. Radford, Andrew, 1981. Transformational Syntax. Cambridge University Press. Semantics Meaning Linguistic analysis focuses on the system of knowledge that makes it possible to speak and understand a language. Knowledge of the meaning of linguistic expressions is part of a speaker's knowledge of his language. Semantics is the study of the meaning of words, phrases and sentences. Semantics deals with the literal (vs. non-literal) meaning of linguistic expressions. Its prime concern is the conventional meaning conveyed by the use of words (or phrases, sentences) of the language rather than what a speaker might want the words to mean on a particular occasion. Ex: [A and B are involved in a nasty argument and A says to B:] 'The door is right behind you.' The literal content of this utterance locates the door behind B. The actual message conveyed is the non-

4 4.1 (1)

(2)

30

literal content 'Leave!', "Get out!" or 'Make use of it!'. It wouldn't be correct to characterize the linguistic (literal) meaning of A's sentence as a command to leave. In another situation the same utterance may convey another meaning, namely a description of the logistics of the room or a suggestion to have the door fixed or closed. Semantics is concerned with literal (invariant) meaning. (3) What is meaning? What is the relationship between a word, phrase or sentence and its meaning? There are different conceptions of meaning and consequently different approaches to semantics, cf. Kempson (1977). 1. Reference. One approach to semantics equates a word's meaning with the word's reference (i.e. the object or objects in the world to which it denotes). In the same way that proper nouns refer to a unique object (cf. the Earth, Chicago), the meaning of the word dog would correspond to the set of entities (dogs) that it picks out in the real word. · There is a problem with words like Pegasus, unicorn, dragon which have no referents in the real world even though they have meaning. · What object would an abstract noun like imagination, patience, talent pick out, or function words like and, what and the? · Though proper names identify an individual, it is not obvious that they have meaning at all: Serena Williams, Noam Chomsky. · Another problem arises with expressions such as the first person to walk on our moon and Neil Armstrong. Both refer to the same person but don't have the same meaning. Cf. also: the morning star, the evening star, Venus. Although these expressions have the same referent, they do not mean the same thing. The husband of Laura Bush doesn't define the term President of US. The impossibility of equating a word's meaning with its referent has led to a distinction between extension and intension. A word's extension corresponds to the set of entities that it picks out in the word, its intension corresponds to its inherent sense, the concept that it evokes. The extension of woman would be a set of real world entities (= all women) while its intension would involve notions like 'female', 'human', 'animal'. The President of the US would have as its extension an individual (at the moment: George W. Bush) but its intension would involve the concept 'elected head of US government'. It has become clear that the question of meaning revolves around a word's inherent sense or intension. 2. Word meanings (intensions) correspond to mental objects or images in the language user's brain. This is an improvement over the referential theory since it is conceivable that one might have a mental image of a unicorn or a dragon even if these entities don't exist in the real world. However: · Mental images aren't abstract enough. An image for the word dog would have to be general enough to include German shephards, greyhounds, poodles, bulldogs, dachshunds, collies, terriers, etc. but it must still exclude foxes, wolves and coyotes. · What would be the 'image' for triangle? (An isosceles, scalene, equilateral or right triangle?) · A single expression may invoke more than one image without itself being ambiguous (a tired child - is he sleeping or screaming?). What type of mental image exists for nitrogen, lecture or because, etc.? 3. Concepts. Meaning must be more abstract than images. One solution attempted

31

in both traditional and modern semantics is to relate meaning to mental concepts of some sort. Since concepts don't have to correspond to objects in the world and need not be images, this approach avoids some of the problems mentioned above. But what is a concept and how is the human conceptual system organized? · Sapir (1921). Language: '... the speech element 'house' is the symbol, first and foremost, not of a single perception, nor even of the notion of a particular object, but of a 'concept', in other words, of a convenient capsule of thought that embraces thousands of distinct experiences and that is ready to take in thousands more.' · Can because, of etc. be analyzed in terms of concepts? · Some concepts may be clear-cut (president, mother), others may be fuzzy (rich, mean). · Concepts may be graded in that they include many exemplars, some of which are more typical of the concept than others. The concept of bird may encompass something like a dictionary definition: 'warm-blooded, egg-laying, feathered vertebrae with forelimbs modified to form wings'. Some creatures are more birdlike than others. Robins and sparrows are better examples of birds than are hummingbirds, ostriches or penguins. Concepts have an internal structure with the best or prototypical exemplars close to the core and less typical members arranged in successively more peripheral regions (= prototype semantics). 4. Componential analysis. A word's intension is equated with an abstract concept consisting of smaller components called semantic features. This approach represents similarities and differences among words with related meanings. It allows the grouping of entities into natural classes. cow animate human male adult + boy + + + man + + + + girl + + woman + + + table -

But what are semantic features? The meaning of blue would contain the feature [+color] and then what - [+blue]? If so, then blue hasn't been broken down into smaller features. The same goes for dog: [+animal, +canine]. 5. Truth conditions. Knowing the meaning of a sentence entails knowing the conditions under which it is true (= possible world semantics): 1. It is raining outside. 2. Sue drives a purple Cadillac. 3. The moon is made of green cheese. These conditions are the truth conditions of a sentence. Sentence meaning is central to this theory; word meaning is characterized as the systematic meaning the words contribute to the meaning of a sentence. Truth conditional semantics doesn't equate meaning with reference or images or concepts. 6. Cognitive science. The meaning of a linguistic expression is a mental representation. The brain is a system that processes mental representations. Meaning must be something that exists in the mind rather than in the world. Jackendoff (1983): We have maintained unrelentingly that word meanings must be

32

treated as internalized mental representations. This rules out in advance an extensional theory of meaning, which identifies the meaning of 'dog' with the set of all dogs (or with the set of all dogs in all possible worlds). Cohen & Murphy (1984): In our view, the path to conceptual clarity lies in basing a theory of concepts on computational theories of knowledge representation rather than on formal semantic or set models. ... Furthermore, we argue that the knowledge representation approach provides a common ground for psychologists, linguists, and researchers in artificial intelligence ... 4.2 (4) Sense Relations The notions of sense and reference (Frege) are central to semantics. a. Constant reference: The expression almost always refers to the same thing (the moon, New York, the United States). There is very little constancy of reference in language. Variable reference: The reference of an expression (your left ear, this transparency) varies according to the circumstances (time, place etc.) in which the expression is used. The context of the expression is needed to fix the reference. b. The sense of an expression is its place in a system of semantic relationships with other expressions in the language. Sense relations. The sense of an expression is its indispensable core of meaning. By virtue of its sense (meaning), a word enters into a variety of semantic relations with other words: synonymy, hyponymy, ambiguity, incompatibility, complementarity. 1. Synonymy is the relationship between two words that have the same sense. Ex: broad, wide; hide, conceal; almost, nearly; liberty, freedom. Two words A and B are synonymous if in every context in which A occurs B may occur without a change in meaning (= substitutability salve veritate (Leibniz' principle)). Synonymy is a special case of hyponymy, i.e. mutual hyponymy. 2. Hyponymy is the sense relation between words such that the meaning of one word is included in the meaning of the other. The meaning of the hyperonym is included in the meaning of the hyponym. The meaning of flower is included in the meaning of daffodil; daffodil is a hyponym of flower. Ex. dog < animal, poodle < dog, carrot < vegetable, oak < tree, table < furniture, scarlet < red. A word A is a hyponym of B if the meaning of B is included in the meaning of A. Each hyponym is set apart from its hyperonyms by at least one further specific feature. 3. Antonymy. Antonyms are words that are opposites with respect to some component of their meaning. There are different types of antonyms. a. Binary antonyms are words that come in pairs and between them exhaust all the relevant possibilities. When one meaning is applicable, then the other cannot be. Ex: male, female; true, false; alive, dead; night, day; smoker, nonsmoker. b. Converses occur when one word describes a relation between two things and another describes the same relation with the two things in the opposite order. Ex. below, above; own, belong to; parent, child; borrow, lend.

(5)

33

c. Co-hyponyms share the same superordinate term (hyperonym): canary, dove, pigeon, duck, flamingo, parrot, pelican, robin, swallow, thrush are co-hyponyms of 'bird'. Pine, elm, ash, weeping willow, sycamore, maple, banyan, oak are cohyponyms of 'tree'. Spring, summer, winter, fall are 'seasons'; hearts, clubs, diamonds, spades are the 'suits of cards' etc. These terms are mutually incompatible (a thing can't be both a parrot and a robin) and together cover all the relevant area. d. Gradable antonyms belong to a scale resulting from a common unit of meaning. They can be used in the comparative. The negative of one member does not necessarily imply the other, not happy is not necessarily sad. <______________________________________________________> cold cool lukewarm warm hot Hot and cold are antonyms because they define the extremities of a scale. Ex: quick, slow; long, short; rich, poor; old, young; shallow, deep; dark, light. Often one member is marked and the other unmarked. The unmarked member is the one used in questions of degree: "How old / *young is he?" - "He is 3 years old / *young." "How expensive / *cheap is it? - "It is very expensive / *cheap." 4. Lexical ambiguity occurs when a single form has more than one meaning. a. Homonymy exists when a single form (written or spoken) has two or more meanings that are not obviously related to each other. Ex: club, pen, bat, pupil, mole; mail/male, meat/meet. b. Polysemy occurs where one form (written or spoken) has multiple meanings which are closely related. Ex: head (person, beer, company), run (animals, water, colors), bright (person, sun), deposit (earth, bank). In reality it is almost impossible to draw a clear line between homonymy and polysemy. c. Conceptual shift. A word often denotes a family of related concepts; the context will decide which concept is meant. 1. Time can be bought for $2.75 or 35 million dollars. 2. There is coffee upstairs. Conceptual differentiation occurs when the meaning of one words determines the meaning option of another word with which it is in construction (Bierwisch 1983). (6) Sense relations among sentences. 1. Two sentences that have the same meaning are paraphrases. a. The cop chased the burglar. - The burglar was chased by the cop. b. I gave the candy to the cat. - I gave the cat the candy. c. It is unfortunate that the team lost. - Unfortunately, the team lost. The sentences have the same truth conditions. It is impossible for one sentence in the pair to be true without the other also being true. Two sentences are paraphrases of each other if and only if they have the same set of entailments (i.e., mutually entail each other) so that when ever one is true the other must be true as well. Two sentences S1 and S2 are synonyms of each other if in every situation in which S1 is true, S2 is also true and vice versa (S1S2). 2. Entailment. The truth of one sentence necessarily implies the truth of another.

34

Paraphrases like those in a.-c. involve mutual entailment since the truth of either sentence guarantees the truth of the other. But entailment can be asymmetrical. d. The park warden killed the bear. - The bear is dead. e. Robin is a man. - Robin is human. f. John boiled an egg. - John cooked an egg. Sentence S1 entails (or includes) S2 if when S1 is true, S2 is also true (S1 S2). Proper inclusion occurs when S1 S2 is valid but not S2 S1. 3. Contradiction exists between sentences when it is the case that if one is true, then the other must be false. g. Charles is asleep. - Charles is awake. h. Peter divorced Sue. - Peter is married to Sue. Two sentences S1 and S2 are contradictory if the truth of S1 implies the falsity of S2 and vice versa (S1 ¬ S2 and S2 ¬ S1). 4. A sentence can be structurally ambiguous. This occurs because the sentence has more than one possible constituent structure: 1. We saw her duck. We saw [S she duck] vs. We saw [NP her duck] or it can be lexically ambiguous if it contains a word that has more than one meaning. 2. The captain corrected the list. (list = inventory or tilt). (7) Presupposition is the relationship between two sentences such that when the presupposing sentence is true the presupposed sentence is true and when the presupposing sentence is false the presupposed sentence is still true. 1. John's brother is a lawyer. 2. John has a brother. 3. John's brother is not a lawyer. 4. John has a brother. If it is false that John's brother is a lawyer, the presupposition that John has a brother still survives. A sentence S1 presupposes a sentence S2 if when S1 is true S2 is true and when S1 is false S2 is still true ( S1 S2 and ¬S1 S2). If we negate an entailing sentence, the entailment fails; but negating a presupposing sentence doesn't affect the presupposition: 5. Sue bought a purple Cadillac. 6. Sue bought a car. 7. Sue didn't buy a Cadillac. 8. §Sue bought a car. 9. The mayor of Berlin will give a speech tonight. 10. There is a mayor of Berlin. 11. The mayor of Berlin will not give a speech tonight. 12. There is a mayor of Berlin. Using a definite description to refer presupposes the existence of the referent. 13. Bill plays first flute in the orchestra. 14. There is a person named Bill. If no such referent exists, the sentence is odd: 15. The King of France is bald. 16. There is a king of France. What is the status of sentence (16)? It is difficult to say that it is false; it is best

35

characterized as inappropriate. If a speaker chooses a definite description like `the King of France' he commits himself to there being a referent to which this expression refers and to the referent being accessible to his conversational partner. A definite description counts as a guarantee that the listener can identify the referent. Factive verbs (like realize, regret) presuppose the truth of their complement clause. 17. Sue realized that Max had lied. 18. Max had lied. 19. Sue regretted eating the whole box of chocolate. 20. Sue ate the whole box of chocolate. (8) An analytic sentence is one that is necessarily true as a result of the senses of the words in it. The truth follows from the meaning relations that exist between the words in it; the truth of the sentence is a function of its linguistic structure. A synthetic sentence is one which is not analytic, but may be either true of false depending on the way the world is. 1. Cats are animals. analytic 2. Cats never live more than 20 years. synthetic 3. Cats are not vegetables. analytic 4. No cat likes water. synthetic Synthetic sentences are potentially informative, whereas analytic sentences are not informative to anyone who already knows the meaning of the words in them. Utterance, sentence and proposition. An utterance is a stretch of speech; it is the use of a certain linguistic expression on a particular occasion. Utterances are physical events. A sentence is not a physical event but an abstract string of words formed according to the rules of a language. (All authentic performances of 'Macbeth' begin with the same sentence, but not with the same utterance.) A proposition is a meaning unit which describes a state of affairs involving a predicate and the referring expressions required by that predicate. 1. Bill studies linguistics. STUDY (Bill, linguistics) Propositions can be true or false (correspond to facts or not). The notion of truth can be used in deciding whether two sentences express different propositions or not. If there is any conceivable set of circumstances in which one sentence is true, while the other is false, we are dealing with different propositions. Predicates and Arguments Argument and predicate. If we delete the referring expressions in the following sentences, we arrive at the predicate of the sentence. Predicates don't refer but require referring expressions. The referring expressions crossed out in the following sentences function as the arguments of the relevant predicate. 1. Our dog bit the postman. bite 2. Sue bought a purple Cadillac. buy 3. He gave the cat smoked salmon. give 4. The airport is near Leipzig. near 5. Bill Gates is wealthy. wealthy The degree of a predicate is the number of arguments it has.

(9)

4.3 (10)

36

6. John sneezed. one-place 7. Bill is foolish. one-place 8. Serena won the match. two-place 9. The marble is under my chair. two-place 10. He sold his car to his best friend. three-place Simple propositions contain just one predicate and its argument(s). 11. Bill gulped. GULP (b) 12. My cat adores my dog. ADORE (c, d) 13. I sent Sue the book. SEND (i, s, b) 14. The cat is under the tree. UNDER (c, t) (11) Semantic/thematic roles. The referents of certain syntactic phrases denote the participants involved in the situation described by the verbal predicate. 1. [The boy AGENT] kicked [the ball THEME]. 2. [The child AGENT] drank [his milk THEME] [with a straw INSTRUMENT]. 3. [We EXPERIENCER] heard [a noise THEME] [in the room LOCATION]. 4. [Sue AGENT] drove [her purple Cadillac THEME] [PATH from Cleveland SOURCE to Chicago GOAL]. 5. [The smoke THEME] swirled [through the room PATH]. 6. [They AGENT] gave [the prize THEME] [to the winner GOAL]. AGENT is the entity performing the action. THEME is the entity that is involved in or affected by the action. INSTRUMENT is the entity used in carrying out the action. EXPERIENCER is the entity (a person) who has a feeling, a perception or a state. LOCATION is a place where a theme is located. PATH is the stretch of space along which an entity moves and can be specified more exactly by a SOURCE as the place from which the entity moves, a ROUTE along which it moves and the GOAL as the place to which it moves. Where do thematic roles come from? How does the grammar ensure that the appropriate thematic role is associated with the correct NP or PP in a sentence? a. Thematic roles originate in word meaning. The sentence the boy kicked the ball has an AGENT and a THEME because the verb kick implies an entity that performs the action it denotes as well as the entity that received this action. kick hear swirl with to V: V: V: P: P: drink V: drive V: give V: from P: through P:

b. Thematic roles are assigned to NPs based on the position of the NP or PP in the syntactic structure. 1. [They AGENT] gave [the prize THEME] [to the winner GOAL]. S

37

NP V they

VP NP PP to the winner

gave the prize

2. [The smoke THEME] swirled [through the room PATH]. S NP V P the smoke swirled through VP PP NP the room

c. Theta Criterion: a. Every theta role of a lexical head must be assigned to a complement of that head. b. Every syntactic complement must be assigned a theta role. 1. *The child drank. / *I hear. / She smiled. 2. *The smoke swirled the chimney. / The smoke swirled up the chimney. 3. *The boy kicked the ball the bat. / The boy kicked the ball with his foot. (12) Compositionality. Up to now we have focused on the meaning conveyed by individual words and the sense relations that exist between related words. The principle of compositionality (Frege) states that the meaning of a complex linguistic expression is the product of the meaning of its component parts and their manner of combination. Idioms are fixed phrases with meanings that are not compositional, i.e. they cannot be inferred from the meanings of the individual words. Each idiom also has a compositional meaning: 1. He will have to eat his words. 2. She let her hair down. 3. He put his foot in his mouth. 4. She always throws her weight around. 5. He kicked the bucket. 6. They bit the grass. Violation of semantic rules. a. The semantic properties of words determine how they combine with other words. Sentences 1. and 2. have the same syntactic structure, but 1. is anomalous. 1. §Colorless green ideas sleep furiously. 2. Dark green leaves rustle furiously. b. Metaphor. Sometimes anomaly is intended and the result can be/is interpreted in a non-literal sense. The literal sense is so unlikely that speakers search for

(13)

(14)

38

another interpretation. Speakers are using language creatively. 3. Our doubts are traitors. (Shakespeare) 4. Walls have ears. (Cervantes) 4.4 References Bierwisch, Manfred, 1983. Semantische und konzeptuelle Repräsentation lexikalischer Einheiten. In: Ruzicka, Rudolf & W. Motsch (eds.), Untersuchungen zur Semantik. Berlin: Akademie Verlag, 61-99. Cohen, Benjamin & G. Murphy, 1984. Models of Concepts. Cognitive Science 8, 27-58. Jackendoff, Ray, 1983. Semantics and Cognition. Cambridge, Mass.: MIT Press Kempson, Ruth, 1977. Semantic Theory. Cambridge University Press. Saeed, John, 1998. Semantics. Oxford: Blackwell. Pragmatics Speech Acts Semantics is the sub-discipline of linguistics that deals with the literal meaning of sentences. Sentences are the grammatical means of encoding the basic semantic unit `proposition'. Pragmatics, on the other hand, is concerned with the meaning of utterances beyond their literal meaning and in the context of their use. Pragmatics traditionally has encompassed three general areas: speech acts, conversational implicatures and presuppositions. Pragmatics is the study not of the grammar itself, but of the use of the products of grammar in actual speech situations. The Descriptive Fallacy: A long philosophical tradition had assumed that sentences were used mainly to describe states of affairs. The purpose of uttering a sentence like 1. Berlin is the capital of Germany. is to describe a particular state of affairs. The sentence has a truth value and can be correlated with a truth condition on its meaning (i.e. it is true if indeed 'Berlin is the capital of Germany'). The Oxford philosopher J.L.Austin (1962) in How to do things with words refutes this idea, which has since become known as the 'descriptive fallacy'. Austin argues that not all utterances in language are used to report (or describe) facts in the world, as for instance the following: - interrogatives Does Sue have two dogs? - imperatives Take that gum out of your mouth! - performatives I confess to you that I stole the last cookie. These utterances don't describe facts and therefore can't be said to have truth values in the same sense that declaratives (Berlin is the capital of Germany) do. Austin (1962): "I want to discuss a kind of utterance which looks like a statement ... and yet is not true or false ... in the first person singular present indicative active ... if a person makes an utterance of this sort we would say that is he doing something rather than merely saying something. ... When I say 'I do' (take this woman to be my lawfully wedded wife), I am not reporting on a marriage, I am indulging in it." The uttering of certain words, in the appropriate circumstances, by and to the appropriate people, constitutes doing something (i.e. marrying, betting, confessing, promising). These are performatives. There are explicit performatives, i.e.

5 5.1 (1)

(2)

(3)

39

sentences that make explicit what one is doing with words. 1. I pronounce you man and wife. (only possible by a clergyman) 2. I sentence you to five years in prison. (only possible by a judge) 3. I bet you five dollars that Serena will win. 4. I challenge you to a match. 5. I dare you to walk through that door. 6. I fine you $500 for possession of marijuana. (only by an authority) 7. I move that we adjourn. 8. I nominate Batman for mayor of Gotham City. 9. I bequeath all my earthly possessions in equal measure to my three siblings. 10. I resign. A test for performatives is the compatibility of the sentence with the phrase I hereby: 11.a. I hereby apologize to you. 11.b. *I hereby know you. 11a. is an act of apologizing; 11b. is odd (semantically anomalous). Not all explicit performatives are in the 1st person, present, indicative, active form: 12. Pedestrians are (hereby) warned to stay off the bridge. 13. You are (hereby) authorized to conduct negotiations in our behalf. (4) In support of his idea, Austin distinguished between performative and constative utterances. A performative utterance is used to perform an act: 14. I promise to repay you tomorrow. In contrast, the statements 15. You promised to repay me tomorrow or 16. I promised to repay you tomorrow only describe a promise but are not themselves promises. These utterances do not simultaneously do what they describe. Consider also the following contrast: A: [John to Bill]: Happy birthday! B: [John sitting at his desk writing a card, Bill approaches] Bill: What are you doing? John: Actually, I am wishing you a happy birthday at the moment! A is an act of congratulations; B is a description of the act. A constative utterance on the other hand makes an assertion. 17. Sue drives a purple Cadillac. Performative utterances can be differentiated from constative utterances by two criteria: 1) Performatives complete an action while constatives report on states of affairs. 2) Performatives are consequently neither true nor false, while constatives are either true or false, since they correspond to facts (a state of affairs in the world). The relevant dimension for judging performatives is whether or not they are appropriate. Dimension for: performatives - appropriate or inappropriate constatives - true or false Austin himself eventually finds the performative/constative distinction problematical and gives up the idea. He realizes that even constative utterances are subject to appropriateness conditions (cf. Bill is innocent) and performatives can turn out to be false (The ball is out). Instead, he proposes that every utterance is a kind of speech act. His theory of speech acts entails that even when there is no

(5)

40

explicit performative verb involved, we recognize - an implicit performance of asserting (Sue drives a purple Cadillac.), - a performance of questioning (Who is Michael Jackson?), - a performance of ordering (Clean up your room!) etc., - a warning (There is a car speeding toward you!). In all these cases we could use an actual performative verb (I state, I ask you, I order you, I warn you...). The illocutionary force of an implicit performative depends on the context of use. For example: the illocutionary act of 'warning' can be accomplished by a statement, an imperative, or a yes-no or wh question: 18.a. There's a bear behind you. b. Run fast! c. Did you know there's a bear behind you? d. What's that bear doing behind you? (6) Proposition, illocution/illocutionary force of an utterance (F(p)). Proposition = p; illocutionary force = F. 19. a. Serena beat Venus. (F1 (p1)) b. Did Serena beat Venus? (F2 (p1)) c. Serena beat Venus? (F2 (p1)) d. Serena, beat Venus! (F3 (p1)) 20. a. You took the last cookie. (F1 (p2)) b. Did you take the last cookie? (F2 (p2)) c. You took the last cookie? (F2 (p2)) d. Take the last cookie! (F3 (p2)) F1 = statement, F2 = question, F3 = request ... The propositional content of p1 is BEAT(serena, venus) and of p2 is TAKE(you, the last cookie). The illocutionary force of an act is derived from the utterance's significance within a conventional system of social interactions or conventions (greeting, promise, threat, accusation, confession, denial etc.). It is the intention of a speaker in using a sentence to make an utterance. The appropriateness conditions of an illocutionary act are conditions that must be fulfilled in the situation in which the act is carried out, if the act is to be carried out properly (or appropriately). - To give an order, the issuer of the order must be in authority over the hearer: [Servant to Queen:] %Open the window! - To be accused of something, the deed must be wrong, i.e. a murder, theft, but not assisting, helping etc. %I condemn you for helping your neighbor. Appropriate conditions are wider in their application than truth conditions. Only declarative sentences may be true or false, but all types of utterances can be considered appropriate or inappropriate. Searle (1969) in Speech Acts asks what conditions are necessary and sufficient for a certain illocutionary act to have been successfully performed in the utterance of a sentence. He illustrates the structure of a speech act with the example `promise'. 21. I promise that I will help you move the couch tomorrow. The most important appropriateness conditions for the act of promising are: 1. Condition of Propositional Content: p predicates a future act A of S. 2. Preparatory Conditions:

(7)

(8)

41

a) H would prefer S's doing A to his not doing A, b) S believes H would prefer his doing A to his not doing A and c) it is not obvious to S and H that S will do A in the normal course of events. 3. Sincerity Condition: S intends to do A. 4. Essential Condition: S intends that his utterance of A will place him under an obligation to do A. (9) Successful Promises: In order for the utterance in 21. to be a successful promise, it must be recognizable as a promise, and the propositional content, preparatory, sincerity and essential conditions must be met: According to the propositional content condition, the speaker must state the intention of carrying out a future act (`I help you'). The preparatory conditions require that the speaker believes that he is able to help and that the hearer wishes to have his help. This condition would be violated if the speaker knew that he was not available tomorrow or that he was incapable of helping. The promise would likewise not succeed if the speaker knew that the hearer did not want his help. For the sincerity condition to hold, the speaker must sincerely intend to help the hearer. This condition would be violated if the speaker had no intention of helping. The essential condition of a promise is that the speaker must intend by the utterance to place himself under an obligation to provide help to the hearer. Appropriateness conditions for making a request: 22. I ask you to please shut the door. 1. Condition of Propositional Content: p predicates a future act A of H. 2. Preparatory Conditions: a) H is capable of doing A, b) S believes H is capable of A and c) it is not obvious to S and H that H will do A in the normal course of events. 3. Sincerity Condition: S wants H to do A. 4. Essential Condition: An attempt to get H to do A. Indirect speech acts. The direct illocution of an utterance is the illocution most directly indicated by a literal reading of the sentence. An indirect illocution is any further illocution the utterance may have. There are conventionalized and non-conventionalized indirect speech acts. a. Conventionalized: Can you pass me the salt? a) the primary illocution is that of a request but b) the secondary illocution is that of a question. b. Non-conventionalized: A: Do you want to go to the movies tonight? B: I have to study for an exam. a) the primary illocution is a rejection, b) the secondary illocution is an assertion. Conventionalized indirect speech acts have a systematic connection to the constitutive rules of the indirect illocution: a. The Preparatory Condition can be asked or asserted: Can you close the door? / You can close the door. = (10.2a). b. The Condition of Propositional Content can be asked or asserted: Will you close

(10)

(11)

42

the door? / Are you going to close the door? You will close the door. (10.1). c. The Sincerity Condition can be asserted: I'd like you to close the door. (10.3) In other cases the speaker asks or asserts reasons for the hearer to do something: [Students in the lecture to Prof. Olsen:] Why don't we stop here? / If we stop now, we would have time for a lunch before our next class. 5.2 (12) Conversational Implicatures Levinson (1983): "The notion of conversational implicature is one of the single most important ideas in pragmatics. [...] The sources of this species of pragmatic inference can be shown to lie outside the organization of language, in some general principles for co-operative interaction, and yet these principles have a pervasive effect upon the structure of language. [...] the notion of implicature [...] provides some explicit account of how it is possible to mean [...] more than what is actually 'said' (i.e. more than what is literally expressed by the conventional sense of the linguistic expressions uttered)." Consider, for example: A: Can you tell me the time? B: Well, the milkman has come." Literally: A: Do you have the ability to tell me the time? B: The milkman came at some time prior to the time of speaking and this fact is relevant at the present. The message actually communicated: A: Do you have the ability to tell me the time of the present moment, as standardly indicated on a watch/clock and if so please do tell me. B: No I don't know the exact time of the present moment, but I can provide some information from which you may be able to deduce the approximate time, namely the milkman has come. The communicative point of the exchange is a request for information and an attempt to supply as much of that information as possible. This is not expressed at all. The gap between what is literally said and what is actually conveyed is so substantial that we cannot expect a semantic theory to account for it." The notion of implicature allows one to assume that expressions in natural languages have simple, invariable senses (literal meaning) but that this stable core of meaning often has a variable, context-specific pragmatic overlay - namely a set of implicatures (utterance meanings). The first of these constitutes the object of the field of semantics built on simple logical principles. Pragmatics, on the other hand, studies the use of language and the inferences that can be derived from an utterance in a particular context. Levinson (1983:100): "... once pragmatic implications of the sort we shall call implicature are taken into account, the apparently radical differences between logic and natural language seem to fade away." The kind of inferences that are called implicatures are, in Grice's sense, intended to be recognized as having been intended. Grice's (1975) theory is essentially a theory about how people use language: There is a set of overarching assumptions that may be formulated as guidelines for the efficient and effective use of language in conversation. These 'maxims' specify what participants must do in order to converse in a maximally efficient, rational,

(13)

(14)

43

co-operative way. The maxims jointly express a general co-operative principle. The co-operative principle: Make your contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the conversation in which you are engaged. Maxims of conversation: 1. The maxim of quality. Make your contribution one that is true: a. Do not say what you believe to be false. b. Do not say that for which you lack adequate evidence. 2. The maxim of quantity. a. Make your contribution as informative as is required for the current purpose. b. Do not make your contribution more informative than is required. 3. The maxim of relevance. Make your contribution relevant. 4. The maxim of manner: Be perspicuous: avoid obscurity, ambiguity; be brief, be orderly. (15) Speakers adhere to these principles in conversation. If they are violated, it is usually the case that at a deeper (non-superficial level) they are being observed. The following exchange violates the maxim of relevance and thus appears to be non co-operative. A. Where's Sue? B: There is a BMW parked outside of Bill's apartment. However, it invokes the inference that the place of the BMW gives some indication about where Sue is. Inferences arise to preserve the assumption that the conversation partner is being co-operative. This is the source of a conversational implicature. The maxims describe rational means for conducting co-operative exchanges of all types (not just linguistic). Consider the following examples taken from Grice concerning a situation in which A and B are fixing a car. a) A needs brake fluid and B, knowing this, hands A the oil instead. Or: A needs the bolts tightened on the steering column of his car and B merely pretends to tighten them. These are violations of quality. b) A needs three bolts and B hands him only one (or passes him 300). Quantity is violated. c) A is in immediate need of a hammer at the moment and B lets him have it an hour later. In this case, relevance is violated. d) A needs a bolt of size 8 and B passes him the bolt of the correct size but in a box that is labeled size 10. B has violated manner. In each of these cases the behavior falls short of some natural notion of cooperation. The linguistic interest in the maxims is that they generate inferences beyond the semantic content of the sentences uttered. Such inferences are conversational implicatures. The term implicature is intended to contrast with the terms logical implication, entailment and consequence which apply to inferences that are derived solely from semantic content. Implicatures are not semantic inferences but inferences based on an assumption about co-operative verbal interaction. Examples of observing the maxims. 1. Relevance. a) A [standing by his stopped car]: I have run out of gasoline. B: There is a garage just around the corner.

(16)

(17)

44

Implicature: The garage is around the corner (not 3 miles away), it is opened, ... b) A: Would you like a cup of coffee? B: Coffee would keep me awake. Implicatures: Yes (before a boring lecture); No (at 11:30pm). 2. Quality (Acts are sincere). a) John has two PhDs. Implicature: I believe this and have evidence that it is true. b) Does his farm contain 400 acres? Implicature: I don't know the answer but want to know it. 3. Quantity (Right proportion; full information needed). a) John has three children. Implicature: John has three and only three children. b) A to B: I would like to introduce you to two of my colleagues, Prof. Jones and Prof. Miller, who both attended your lecture last night. Prof. Jones found it very interesting. Implicature: Prof. Miller didn't find it interesting. c) A: How expensive is an IBM laptop? B: About three thousand dollars. / Two thousand eight hundred and seventy-five dollars and ninety-five cents plus tax and minus a 5% discount. d) Some linguists are very interesting people. Implicature: Not all linguists are interesting. 4. Manner (Clarity of purpose). a) Sue plays tennis and Jim skis = Jim skis and Sue plays tennis. b) The highway patrol drove up and Sue slowed down to 65 mph. Implicature: This is the reason for Sue's slowing down. c) Sue read the newspaper and had dinner Implicature: in this order. vs. Sue had dinner and read the newspaper. Implicatures are triggered by the assumption that the maxims are being observed. They arise wherever features of the context do not actually block them. (18) Speakers can deliberately (overtly and blatantly) violate (flout: or exploit) the maxims: 1. Quality. a) A [in Texas in July]: What is the weather like today? B: It's probably cool as usual. Implicature: irony. b) John is a machine. Implicature: metaphor; John has the (positive) property of being extremely efficient or the (negative) quality of lacking flexibility, emotion etc. 2. Quantity. a) Polonius: What do you read, my lord? Hamlet: Words, words, words. (Hamlet is feigning insanity). b) Politics is politics. Either John will make it or he won't. It's not over until it's over. If he does, he does. Implicatures arise because tautologies (logical truths) are non-informative; if speaker is intends to be informative, then some inference must be made. 3. Relevance. a) A: John is divorcing Mary. B: Nice day today, isn't it? Implicature: I don't want to talk about it. b) Johnny to his brother: Lets play a video game! Mother: Have you two done your homework? Implicature: Reminder. 4. Manner. a) A: I saw Sue in the opera last night with a good-looking man. Implicature: The man is not Sue's husband/son etc. b) A: [in a letter of recommendation]: You will be fortunate indeed if you can get Henry to work for you. Implicature: Since the sentence is ambiguous, it is possible

45

to take it in its negative sense. c) A: Let us buy the children something. B: Okay, but I veto i-c-e c-r-e-a-m. Implicature: The children are not supposed to understand the message (i.e. the spelled word). (19) Grice's schema for calculating an implicature. (How one arrives from the literal meaning of the utterance at the sense intended by using the co-operative principle and the maxims.) a. S has said that p. b. There is no reason to think S is not observing the maxims, or at least the cooperative principle. c. In order for S to say that p and still observe the maxims or the co-operative principle, S must think that q. d. S must know that it is mutual knowledge that q must be supposed if S is to be taken as co-operating. e. S has done nothing to stop me thinking that q. f. Therefore S intends me to think that q, and in saying that p has implicated q. Characteristic properties of implicatures: a. They are cancelable: John has two sons (the implicature: he has two and only two sons). But the implicature can be cancelled: I mean at least two sons; he may have more. An implicature disappears when it is clear from the context that such an inference could not have been intended as part of the utterance's full communicative import. Ex.: [Ticket seller at a cinema showing a film that is restricted]: Is your son 12 years old? B: Yes. (B's son is 14). b. Implicatures are non-detachable. An implicature is attached to the semantic content of what is said, not to the linguistic form of the utterance and therefore it cannot be 'detached' from the utterance simply by exchanging one word for a synonym. Ex.: [John has made Bill angry with his stupidity. Bill says]: John is a genius (ironical). The synonymous expressions in (i)-(iii) all carry the same implicature in this context: i. John is a mental prodigy, ii. John is an fabulous intellect, iii. John has an enormous amount of gray matter ... c. Implicatures are calculable, as demonstrated in (19) above. d. Implicatures are non-conventional, i.e. not part of the conventional meaning of the linguistic expression. The fact that they are cancelable and non-detachable shows this. One needs to know the literal meaning of a sentence before one can calculate its implicature in a context, hence the implicature cannot be part of the conventional meaning. References Austin, J.L., 1962. How To Do Things With Words. Oxford: Oxford University Press. Grice, H.P., 1975. Logic and Conversation. In: P. Cole & J.L. Morgan (eds.), Syntax and Semantics 3: Speech Acts., New York: Academic Press, 41-58 Levinson, Stephen, 1983. Pragmatics. Cambridge University Press. Searle, J.R., 1969. Speech Acts. Cambridge: Cambridge University Press.[Ch. 1-3] Searle, J.R., 1975. Indirect Speech Acts. In: P. Cole & J.L. Morgan (eds.), Syntax and Semantics. Vol. 3, New York: Academic Press.

(20)

5.3

46

6 6.1. (1)

Phonology Evolution of the Vocal Tract Human language is a comparatively recent evolutionary phenomenon that added two powerful innovations to older systems: speech and syntax. It may have reached its present form 100,000 years ago. Chimpanzees, gorillas and other non-human primates can comprehend and, in some cases, even use words, as attempts to teach apes human language have demonstrated. Charles Darwin characterized nature as a "miserly opportunist". What he meant was that the process of evolution always makes use of old parts, modifying them to perform new functions. The biological bases of the newest, unique components of human language, the tongue and mouth as well as the brain mechanisms that regulate speech production and syntax, evolved from the tongues, mouths and brains of archaic hominids. This means that organs that were originally designed for breathing air and swallowing food were adapted to produce human speech. There are selective advantages to the use of speech. Sign language can be used for communication, but the signer's hands are not free for other tasks at the same time and the gestures can only be seen under restricted conditions. Vocal communication, on the other hand, preserves the advantages gained by the evolution of upright locomotion in the earliest hominids approximately four million years ago. Present-day apes have the anatomical organs for generating a subset of the sounds of human speech. This could mean that vocal language in a limited form may date back to them. The particular sounds of human speech - which as a whole are out of the range of apes' abilities - are an important component of human linguistic. The sounds of human speech allow us to transmit phonetic 'segments' at an extremely rapid rate, up to 25 segments per second. In contrast, human beings can only identify nonspeech sounds at rates of approximately 7 to 9 units per second. To illustrate the importance of this fact, a short sentence that contains about 50 speech sounds can be uttered in two seconds and human listeners have no difficulty in understanding what has been spoken. If the sentence were transmitted at the non-speech rate, however, it would take so long that a listener might forget the beginning of the sentence before hearing its end. Rapid speech, therefore, requires a special set of brain mechanisms that decode speech signals in a task-specific way. The speech sounds that only Homo sapiens can produce are also more reliable than the range of sounds accessible to apes. Only human beings can produce non-nasal sounds and vowels like [i] and [u]. These sounds are less susceptible to perceptual confusion than the sounds that other primates can make. The physiology of speech. The supralaryngeal vocal tract (SLV) necessary for human speech has some odd properties compared to SLV of various extinct hominids. Human beings are more liable to choke when eating or drinking than other mammals: Solid objects or liquid can fall into the human larynx, blocking the pathway to the lungs. However, the human vocal tract is very well suited to the production of speech. The tongue can produce abrupt changes in the cross-sectional area of the vocal tract necessary to articulate the vowel sounds [i], [u] and [a] as well as velar consonants such as [g] and [k]. The fact that the velum of the palate

(2)

(3)

(4)

(5)

47

can seal off the entrance to the nasal cavity also allows humans to produce nonnasal speech. Although these slight additions have reduced biological fitness, they play an important role in making human speech rapid and efficient. (6) Humans have innate brain mechanisms that are specialized for speech perception and production. The articulatory maneuvers that underlie speech (rapid coordination of the tongue, lips, velum, larynx and lungs) are extremely complex. Children do not attain adult levels of precision in the articulation of vowels until around the age of 10. In decoding the speech signal, we unconsciously map patterns of formant frequencies into discrete phonetic categories. One pattern of formant frequencies will be heard as a [b], another as a [p], etc. The speech detectors, after hearing 1/100 of a second of the start of a monosyllabic word like bee, will 'hear' the entire syllable. The larynx changes the flow of air from the lungs into periodic puffs. The puffs are produced by the rapid opening and closing of the vocal cords. The supralaryngeal vocal tract acts like the pipes of an organ in letting acoustic energy through at certain frequencies. The frequencies at which maximum acoustic energy flows through the vocal tract are 'formant frequencies'. During speech the shape of the vocal tract is continually changing, thereby generating different formant frequency patterns. Human listeners have to calculate the formant frequency pattern from the acoustic signal and estimate the probable length of a speaker's supralaryngeal airway in order to assign a particular formant frequency pattern to a particular speech sound. The length of the vocal tract of a young child is half the length of an adult. The different lengths have different formant frequencies. A short tract will have higher formant frequencies than a long one (cf. a piccolo to a bassoon). If we interpreted the formant sounds of speech like the notes produced by woodwind instruments, then the speech sounds of people who had different vocal tract lengths would not have the same phonetic value. For example, the word bit spoken by a large adult male speaker can have the same formant frequency pattern as the word bet produced by a smaller male. Yet we 'hear' the large person's bit as bit rather than bet. (8) The selective advantage of the human supralaryngeal vocal tract versus that of a chimpanzee, Australopithecine or Neanderthal for speech is that it can produce the supervowel [i] and speech that isn't nasalized. The supervowels are optimal signals for estimating the length of the vocal tract producing the speech signal that is being decoded. Other sounds work too, but [i] works best and [u] next best. All human languages incorporate these vowels. The ability to produce sounds that are not nasalized facilitates extracting the formant frequencies from the acoustic speech signal. Nasalization increases the error rate for vowel identification from 5% to 3050%. In contrast, non-human primates are unable to produce the voluntarily muscular maneuvers that underlie human speech. However, computer-modeling shows that a chimpanzee could produce a subset of human speech sounds such as the vowels [I], [e], [ae], [U] and the consonants [t], [d], [b], [p]. That is, they could speak nasalized versions of food or bit. 6.2 (9) The Phonological Base: Phonemes The study of speech sounds is the object of two related linguistic sub-disciplines: phonetics and phonology. Phonetics studies the physical (acoustic) aspects of the

(7)

48

speech signal, while phonology is concerned with the patterns of the sound systems of natural languages as they belong to the linguistic competence of native speakers. (10) The phonological system of a language makes use of a system of contrasts. The sounds that make up the contrasts of the system are termed phonemes. A phoneme is the term used to designate an abstract class of non-contrastive sounds comprising a number of variant pronunciations or allophones of the phoneme. Ex: In English the voiceless stops /p, t, k/ are pronounced with aspiration in the words in (1a) but without aspiration following the /s/ sound in the words in (1b): 1. a. pan tan can b. span Stan scan

The lack of aspiration following /s/ is predictable; that is, it never serves to distinguish two words in English. Therefore, the aspirated variants of /p, t, k/ do not contrast with the unaspirated variants but rather form variants belong to the same phoneme. In Hindi, on the other hand, the contrast between aspirated and unaspirated stops is phonemic, since it differentiates between words: pal and phal are independent words. They constitute a minimal pair in (2a). 2. a. pal `take care of' b. tan `mode of singing' c. kan `ear' phal `edge of knife' than `roll of cloth' khan `mime'

(11)

The English forms in (3) also form a minimal pair: They contrast in one sound segment only and the difference of this one sound distinguishes different words. The minimal contrast in (3) therefore demonstrates that /r/ and /l/ are two different phonemes in English. 3. a. right b. light

The lack of contrast between (1a) and (1b), on the other hand, indicates that unaspirated and aspirated stops are simply allophones of the same phoneme. Aspirated and unaspirated voiceless stops in English are said to be in complementary distribution: After [s] voiceless stops are unaspirated, otherwise they are aspirated. 4. English /p/ [p=] [ph] 5. Hindi: /p=/ [p=] 6.3 (12) /ph/ [ph] [t=] /t=/ [t=] /t/ [th] /th/ [th] /k/ [k=] /k=/ [k=] [kh] /kh/ [kh] phoneme phone phoneme phone

Articulation of Consonants Manner of Articulation. The source of speech sounds is the air passing from the lungs through the larynx and on into the cavities in the oral tract. These cavities can be modified in shape by the different positions of the articulators (the tongue, lips,

49

velum etc.). The way the shape of the cavity changes influences the way the air in it vibrates giving rise to different sounds. Sonorants involve a high degree of resonance, obstruents are characterized by much less resonance. Obstruents (plosives, fricatives and affricates) are produced when the air passing through the mouth is obstructed by a narrowing or closure in the oral tract resulting in friction or in a sudden release of sound. Plosives are produced by a complete closure blocking the air momentarily and then releasing it abruptly. Plosives are also called stops. A fricative is made by an incomplete closure (or stricture) which produces friction as the air is forced through it. Fricatives are also termed continuants. An affricate combines a stop with a following continuant but lasts only as long as a single fricative. Sonorants result when the buccal cavity (the throat, mouth and nasal passage) resonate. Nasal stops are sonorant consonants formed by lowering the velum which allows the air to pass through the nasal passage and at the same time creating a blockage of air in the oral cavity. Approximants are frictionless continuants and are therefore also sonorant. Approximants don't involve a blockage or obstruction of air in the mouth as is the case with stops and fricatives. This results in glides ([w] and [j]) and liquids ([l] and [r]). Glides are basically high vowels that don't function as the nucleus of a syllable and are, in this sense, somewhat consonant-like. Liquids consist of lateral (/l/ sounds) and rhotic (/r/) sounds. In the production of laterals, the tongue obstructs the passage of air through the middle of the mouth by pressing against the alveolar region behind the teeth. The air stream escapes without friction laterally, i.e. along the sides of the oral cavity. (13) Place of Articulation. The air stream used in producing speech sounds comes from the lungs and passes through the larynx and the vocal cords. Some of the components of the sound so produced are filtered out by certain configurations of the vocal tract, while other components are simultaneously amplified by them. Labial and labiodental sounds are produced when the lips or the lips and the teeth form a stricture; velar sounds are produced by a construction with the tongue against the soft palate (velum); glottal sounds (like the glottal fricative [h] and the glottal stop [§]) are produced by a constriction of the vocal cords. Coronal sounds are made by raising the front part of the tongue against the area extending from the teeth to the hard palate. (14) Voicing. The larynx contains vocal cords which are attached to an immovable part of the larynx (the thyroid cartilage) at one end and to movable parts (the arytenoid cartilages) at the other. When the vocal cords are brought together the outgoing air causes them to vibrate creating a voiced sound. Basic consonants of English

(15)

bilabial

labiodental

dental/alveolar/a-p. palatal

velar

glottal

50

stop p b nasal m flap fricative approximant lateral

t d n f v > s z · ¥ < l j

kg õ

§

h

6.4 (16)

Articulation of Vowels Vibrations from the vocal cords produce sounds of different frequencies which - in the absence of a stricture - resonate in the vocal tract. Vowels are thus sonorants. As the oral cavity changes shape as a result of the position of the tongue and the shape of the lips, the way the air in it is amplified will also change. A vowel is thus a mixture of sounds at different amplified frequencies. Vowel quality is determined by the articulators: The most important aspects are tongue height (high, mid, low), frontness (front, central, back), lip rounding (rounding, unrounded) and advanced tongue root (tense, lax). High vowels are produced by raising the body of the tongue beyond the mid position characteristic of the schwa, low vowels entail lowering the tongue and back vowels involve retraction of the tongue body. Round vowels are produced by rounding the lips, unrounded vowels lack this rounding. Tense vowels result when the tongue root is pushed forward from its normal resting position; lax vowels are produced without the advanced tongue root. Vowels can differ in quantity: A long vowel is about twice as long as a short vowel. Quantity also distinguishes monophthongs (simple long or short vowels) from diphthongs which are two vowels pronounced in the same amount of time as a long monophthong. A diphthong basically comprises two half vowels, the first undergoing a rapid transition into the second. One of the two half vowels will dominate over the other. If the dominant one is the first (cf. English [aI]), the diphthong is termed falling, otherwise it is rising.

The US vowel system (after Spencer (1996:32). Monophthongs i I e g æ a
Diphthongs

u c o ]

aj

aw (æw)

]j

51

Transcriptions: 1. pit pot pea pay
Po

2. 3.

/pIt/ /pat/ /pi/ /pe/
/po/

pet putt poo

/pgt/ pat /p t/ put /pu/ paw

/pæt/ /pt/ /p]/ Papa

/papa/

pie /paj/ point /p]jnt/ pow /paw/

6.5 (18)

The Structure of the Syllable Phonotactic constraints characterize the restrictions on the way sounds are arranged in a particular language. They generally apply at the level of syllable structure. Many phonological processes also operate at the level of the syllable or some constituent of the syllable. Therefore, an understanding of syllable structure is essential for understanding the phonological organization of a language. For example, a rule might affect a consonant but only at the beginning or at the end of a syllable (cf. Auslautverhärtung which disallows voiced obstruents at the end of a syllable in German). The structure of the syllable is hierarchical. It is made up of an onset and a rhyme. The rhyme includes a nucleus and a coda, cf. the structure of strengths, which is the most complex syllable type of English: 1. Onset Syllable Rhyme Nucleus * s t r g õ > s Coda ' * (

(19)

(20) Possible syllable types in English 2. *
R

*
R

*
R

'(
O R

*
R

'(
O R

*
N

*
N

*
N

*
N X

*
X X

*(
N Co

*
N

'(
Co X X

*
X

*(
X X

*(
X X

* * (* *
X X

* * *
X

a

a

a

I

k

aa

t

k

a

t

Neither the onset nor the coda is an obligatory part of the syllable, cf. the first three

52

diagrams in 1. A short vowel or single constituent occupies a single timing slot (= X). A long vowel or a diphthong occupy two timing slots. The different levels in a hierarchial representation are called tiers. The lowest level is the timing tier. The sound qualities represented by the symbols /a, I, k, t/ are melodies. (21) A syllable with a short vowel and no coda (CV) is called light; one with a long vowel or diphthong in the nucleus or a coda (or both) is heavy (CV:, CVV, CVVC CVC or CVCC). Another way of expressing this is: A syllable is heavy if it has a branching rhyme, as is the case with all the diagrams in 2. but the first. The sonority scale. Whereas 3a are possible syllables in English, 3b aren't. 3. a. play pray puce dry dwell b. *lpeI rpeI jpu:s rdaI wdgl

(22)

It is a common phenomenon in languages that the order obstruent + sonorant is possible in the onset of a syllable but the opposite ordering sonorant + obstruent is not well-formed. In the same manner, 4a are possible syllables in English, but 4b aren't. 4. a. lamp lens sink help cold puls carp word horse b. *lapm lezn sikn hepl codl pusl capr wodr hosr

Again, it is common that the opposite order sonorant + obstruent is possible in the coda of a syllable but the ordering obstruent + sonorant is not. Why do we find this pattern? Sonority is a matter of degree. The scale in 5. gives the relative sonority for sounds ranging from the most sonorous (vowels) to the least sonorant (plosives): 5. Sonority Scale: plosives < fricatives/affricates < nasals < liquids < glides < vowels

The relative sonority of sounds is reflected in the structure of the syllable: The nucleus of a syllable is always the most sonorant element, while the onset tends to increase in sonority towards the nucleus and the coda tends to decrease in sonority away from the nucleus. This is illustrated for the syllable blimb in 6. The generalization that captures this regularity is stated according to Selkirk (1984) in 7.: 6. L O b 7. l I m V N O p

Sonority Sequencing Generalization = SSG In any syllable, there is a segment constituting a sonority peak that is preceded and/or followed by a sequence of segments with progressively decreasing sonority values.

53

The sonority scale together with the SSG rules out the ill-formed syllables of 3b and 4b. 6.6 (23) Phonological Processes An example of a phonological process is the devoicing of voiced fricatives in fast speech English. The first column in 16. represents careful speech, the second more natural quick speech: 1. a. five past b. love to go c. as well as can be d. has to go e. loathe to go [faIvpæst] [faIfpæst] [lvtcgou] [lftcgou] [czwelczkcnbI][czwelcskcnbI] [hæztcgou] [hæstcgou] [loutcgou] [loutcgou]

Apparently a voiced fricative is changed to a voiceless fricative immediately before another voiceless consonant. This is a process of assimilating the voice feature of the preceding fricative to the voice feature of the following consonant. The process involved can be captured by the following rule: 2. voiced fricative voiceless / ____ voiceless

The `____' following the slanted bar in 2. indicates the environment in which the sound change occurs: The change is conditioned by an immediately following voiceless sound. 6.7 References: Liebermann, Philip, 1992. On the Evolution of Human Languages. In: J.A. Hawkins & M. Gell-Mann. The Evolution of Human Languages. AddisonWesley, 21-47. Selkirk, Elisabeth, 1984. On the major class features and syllable theory. In: Aronoff, M. & R. Oehrle (eds.), Language Sound Structure: Studies in Phonology Presented to Morris Halle by his Teacher and Students. Cambridge: MIT Press. Spencer, Andrew, 1996. Phonology. Oxford: Blackwell. Neurolinguistics: Structure of the Brain, Aphasia Neurolinguistics is the study of the neurological basis for language. The human brain is composed of nerve cells (neurons) that are the basic information processing units of the nervous system. The brain contains about 10 billion neurons that are organized into networks of extremely high complexity: Each neuron can be directly linked up with over 10,000 other neurons. The structure of the brain. The brain apparently evolved from the bottom up: The lower brain structures are shared by almost all animals. They control the automatic and unconscious aspects of our behavior such as breathing, circulation, heartbeat and digestion - i.e. processes essential to the survival of all animals. Farther from the spinal cord, however, there are structures that have developed differently in

7 (1)

(2)

54

different species. At the highest level of the brain (i.e. at the cerebral cortex), the differences are the greatest. Reptiles and amphibians have no cortex at all and progression from lower to higher mammals is marked by an increase in the proportion of cortex to total amount of brain tissue. In lower mammals most of the cortex is devoted to sensory and motor functions. With chimpanzees and homo sapiens a large part of the cortex is not taken up by the senses. This area is the association cortex. The human cortex is a gray wrinkled mass (gray matter). Its foldings allow the large amount of cortical matter to be compressed into the skull. Up to 65% of the cortex is hidden within its folds. Underneath billions of fibers (or white matter) interconnect the neurons of the cortex. The brain is divided into two cerebral hemispheres that are almost completely separate. The two hemispheres are connected in the middle by the corpus callosum which allows them to exchange information. This pathway between the two hemispheres consists of two million fibers connecting the cells of the left hemisphere to comparable areas of the right hemisphere. Anatomically each hemisphere appears to be the direct counterpart of the other and the basic functions of motor and sense are evenly divided between them, although the organization is contralateral, that is each hemisphere is responsible for the opposite half of the body. The left hemisphere controls the right side of the body and the right hemisphere the left side. (3) Bilateral duplication or cerebral dominance? a. Marc Dax, a practical doctor working in Maine, reported in a lecture to the medical association in Montpellier (capital city of Maine) in 1836 that there must be a connection between the side of brain that is damaged and the loss of language. More than 40 patients that he had treated for aphasia had damage to the left brain and none to the right side of the brain. His theory was that each hemisphere is specialized for unique functions and language is controlled by the left hemisphere. At this point no one took particular notice of his observation. b. Soon the location of language in the brain gained interest due to the work of Paul Broca. Between 1861-64 he had examined patients with aphasia. Autopsies on patients with aphasia that had died showed lesions in the left frontal lobe. Furthermore, aphasic patients still living were hemiplegic on the right side. Lesions in the right hemisphere didn't result in aphasia. Three of Broca's ideas became important: a) Language can be disrupted independently from other cognitive abilities, b) language is located in a specific area of the brain. Broca's studies contained detailed reports on the exact locality of the damage as well as descriptions of the language deficit (or: aphasia) such damage produced. c) Broca hypothesized that the hemisphere that controls language also accounts for the preference of the opposite hand. Cerebral localization: Specific aspects of our cognitive ability have identifiable locations in the brain. One needs to be cautious in drawing definite conclusions about the localization of specific abilities, however. The brain is made up of an intricate interconnection of neurons dependent on a complex system of blood supply. Most brain functions are interdependent. Generally one can state that analytic processes are typical of the left hemisphere, in particular those concerned with language production and comprehension. The left hemisphere seems to be best at processing serial or sequential information. The right hemisphere, on the other hand, is particularly apt at processing spacial information. It seems to process information in parallel and as a whole.

55

(4)

Language centers. The left hemisphere of the brain normally contains the general locations of language functions. 1. Anterior speech cortex or Broca's area. 2. Posterior speech cortex or Wernicke's area. 3. The motor cortex controlling the movement of the muscles lies close to Broca's area. This part of the motor cortex controls the articulatory muscles of the face, jaw, tongue and larynx. 4. The arcuate fasciculus is a bundle of nerve fibers that forms a crucial connection between Wernicke's and Broca's area. Aphasia is an impairment of language function due to localized cerebral damage. Language disorders are often interrelated. Consequently, the classification of types of aphasia is normally based on the primary symptoms. Aphasia is almost always the result of injury to the left hemisphere (= left-hemisphere dominance for language). Deaf signers with damage to the left hemisphere show aphasia for sign language similar to the language breakdown in hearing aphasics. 1. Broca's aphasia (agrammaticism) is characterized by reduced speech with slow, effortful, distorted articulation. What is said consists almost entirely of content words (nouns and verbs). Functional elements (articles, prepositions, pronouns, inflections, conjunctions) are missing. Comprehension is generally good. Example: Yes - ha - Monday - ah - Dad - and Dad - ah - Hospital - an ah - Wednesday - Wednesday - nine o'clock and ah Thursday - ten o'clock ah doctors - two ah doctors and - ah - teeth yah. And a doctor - ah girl - and gums, and I. Testing reveals considerable impairments when comprehension depends on syntactic structure. Broca's aphasics tend to interpret a sentence like (1a) correctly, but not (1b). The meaning (independently of the sentence structure) guides the interpretation of 'the boy' as the agent of the action in (1a) but no such semantic clue as to which entity is the agent of the action is given in (1b). The meaning of the sentence depends on the correct interpretation of the functional morphemes indicating passive. 1. a. The ball was kicked by the boy. b. The girl was kicked by the boy. Broca's aphasics also have difficulty in judging the grammaticality of the sentences in (2): 2. a. The boy ate it up. b. *The boy ate up it. c. *Boy ate it up. 2. Wernicke's aphasia contains fluent speech that often contains lexical errors (word substitutions) and is largely devoid of sense. These aphasics have serious comprehension problems. Example: I felt worse because I can no longer keep in mind from the mind of the minds to keep me from mind and up to the ear which can be to find among ourselves. Some of the structure is well-formed but some of it contains unrelated fragments. 3. Conduction aphasia results from damage to the arcuate fasciculus. There are no articulation problems and comprehension is good. However, the task of repeating a word or phrase is difficult. Apparently what is heard and understood cannot be transferred to the speech production area. 4. Mixed transcortical aphasia (echolalia) occurs when the entire speech area (Broca's area and Wernike's area as well as the connections between them) remains

(5)

56

intact but are isolated from the rest of the conceptual ability. Case study of HCEM who was in a nursing home due to pre-senile dementia and acute chronic alcoholism. She was totally unable to take care of herself. She did not communicate with anybody and had no spontaneous speech. She couldn't even find her own room on the floor. She repeated almost everything addressed to her. HW: Good morning. HCEM: Good morning. HW: How are you this morning? HCEM: How are you this morning? HW: Did you have a nice day yesterday? HCEM: Did you have a nice day yesterday? HW: What did you have for breakfast? HCEM: What did you have for breakfast? The following examples demonstrate an interesting phenomenon, namely that her echolalic mechanism was coupled with a 'grammatical filter': When confronted with phonologically or syntactically incorrect sentences, she responded with grammatically correct ones more than half of the time. However, semantically mistakes were not corrected. HW: There are two book on the table. HCEM: There are few books on the table. HW: She buy a dress yesterday. HCEM: She bought a dress yesterday. HW: This is a gold ling. HCEM: This is a gold ring. Semantic anomalies: HW: The pencil smells the flowers. HCEM: The pencil smells the flowers. HW: I ate a car for dinner. HCEM: I ate a car for dinner. "The reason echolalia occurs in a specific conversation situation is owing to the fact that by establishing eye contact with a patient who is otherwise disoriented, one captures the attention of the patient and impresses upon her the intention of communication. Since the patient has an intact auditory mechanism for the perception of language, she is able to decode the incoming stimuli as linguistic. But since she has no access to cognition or to the creative aspects of language use, the decoded stimuli fail to evoke any response natural to the context of communication. Subsequently, the decoded stimuli become, themselves, encoded output. When syntactically and phonologically wrong stimuli are presented to the patient, they impinge upon the automatic parts of the grammar stored in the language area; these aspects of the grammar cause the erroneous input to conform to the grammar which exercises a filtering function without volitional or conscious effort on the part of the patient." (Whitacker 1976) (6) Lateralization refers to the fact that the two hemispheres show functional distinctness with respect to higher cognitive functions. The right brain excels in tasks which require an overall assessment of complex patterns (recognizing faces, melodies, and intonation patterns as well as handling spatial orientation etc.). The left hemisphere is superior in analytic tasks (arithmetic, language, rhythmic perception, temporal-order judgments etc). The two hemispheres differ in the manner in which they treat incoming stimuli, the right hemisphere processes stimuli holistically (as wholes) and the left analytically (by parts).

57

In an experiment conceived to discover more about the lateralization of the brain, Doreen Kimura of the Montreal Neurological Institute exposed 3 to 10 dots to each visual half-field for 80 milliseconds. Subjects exhibited a left visual field (right hemisphere) superiority in guessing the number of dots. The brevity of the exposure time prevented the subjects from counting the dots. This supports the assumption that the right hemisphere (associated with the left visual field) is superior at grasping the whole without a complete analysis of its parts. Patients with damage to the right brain can have difficulty in recognizing familiar faces or in differentiating the faces of two different people. Aphasics (resulting mostly from damage to the left hemisphere) can still sing, a function which is anchored in the right hemisphere. One patient is described as follows: "er erlitt einen Anfall einer schweren Krankheit, die zur Lähmung der gesamten rechten Körperhälfte und zu einem vollständigen Verlust der Sprache führte. Er kann bestimmte Kirchenlieder, die er vor seiner Erkrankung gelernt hat, so klar und deutlich singen wie jede gesunde Person. Dennoch ist dieser Mann stumm ..." (7) Charting the brain. Surgeons operating on the brain must determine where the speech centers are located so that they won't be damaged during surgery. One way of determining their location is by stimulating different areas of the brain electrically. Since the brain has no receptors for pain, the stimulation can take place while the patient is conscious. The stimuli cause involuntary movements of the arms and legs or trigger different sensations such as sight, sound, smell or taste, when the appropriate centers are stimulated. During the identification of the language centers the patient and surgeon are separated by a curtain and a third person observes the patient's reaction during stimulation. A stimulus to a language area will disrupt the linguistic function (aphasic arrest). The patient is shown a series of pictures which he must describe while an electrode is moved over the surface of the brain. This procedure takes about 15 minutes. The result is a map of the speech centers of the brain. Another means of determining the speech centers in the brain is the Wada Test invented by Juhn Wada. Several days before a scheduled operation, the brain surgeon anesthetizes one hemisphere and then the other by injecting natriumamytal into the main artery of the hemisphere. This barbiturate deadens the hemisphere. During the procedure, the patient is conscious and is lying on his back. He is asked to spread his arms out and to begin counting from 1 to 20. As the natrium-amytal takes effect the arm contralateral to the numbed hemisphere drops. This indicates to the surgeon that the barbiturate has arrived into the correct hemisphere. If the hemisphere is effected that controls language, the counting also stops and the patient is without language for the length of the anesthetic (2-5 minutes). Split-brain patients are epileptic patients who have had their corpus callosum partially severed as means of alleviating seizures. Seizures arise as a result of an excess amount of electrical energy in a focal area of the brain and spread to other areas. Cutting the corpus callosum prohibits the seizure from spreading in extreme cases to the healthy hemisphere. This operation doesn't seem to alter the overall behavior of the patient to any noticeable degree. Split-brain patients have given scientists an opportunity to study the capacities of one hemisphere apart from the other (albeit on the basis of a mafunctioning brain). In one experiment, patients were blindfolded and then given objects to feel (which they couldn't see). If the object was felt with the right hand that is connected neurologically to the left

(8)

58

hemisphere, the patient could name the object. If felt with the left hand, the patient wasn't able to name what was felt. The information apparently doesn't arrive at the speech centers of the left hemisphere. However, the same patient ­ still blindfolded ­ was able to identify the objects felt with the left hand from a larger group of objects. Studies on split-brain patients have indicated that the two hemispheres of the brain process information in different ways. When patients were shown pictures of objects that could be classified according to their function as well as their physical appearance, the left hemisphere classified them according to their function, while the right classified them according to their appearance. Another curious phenomenon of interest was found with split-brain patients. When shown a picture of a composite face made up of the right side of the face of a child and the left side of the face of a woman (a chimaera) and asked to identify the face, a patient who had perceived the chimaera first with her left hemisphere named the child. When the face was perceived only by her right hemisphere, however, the patient pointed to the woman. The opposite hemisphere seems to be blind to the part of the picture seen by the other hemisphere; however, the single hemispheres apparently visualize the picture as a whole and not as a half. This is termed `visual completion'. (9) One means of studying the healthy mind is dichotic listening. Patients hear via a headset different sounds directed over a distinct channel to the right and left ear. There is a right ear advantage (REA) for speech sounds that travel from the right ear directly into the left hemisphere. For non-speech sounds a left ear advantage was found. The advantage manifests itself in that the sounds are recognized more quickly and more clearly. Other more recently developed methods to measure the activity of the brain noninvasively are the CAT and PET scans, nuclear magnetic resonance (NMR), electroencephalogram (EEG) and the event-related potential (ERP). When the brain is subjected to a stimulus, specific changes occur that are ordinarily indistinguishable from the general background activity taking place in the brain. In order to expose the activity related to a specific stimulus (event-related activity), a computer measures the waves that occur in many cases of the same stimulus and computes an average. This processes abstracts away from the irrelevant electrical background activity and captures the activity related to a specific event. The evoked potential consists of a wave of positive and negative variations plotted against a base line. Each potential is defined according to its amplitude and latency. The exact form of the ERP depends on the type of stimulus. Linguistic stimuli result in higher amplitudes in the left hemisphere, non-linguistic produce higher amplitudes in the right hemisphere. The degree of hemispheric specialization for language varies among individuals. The Wada test offers the basis for the most familiar studies on the relation between left- and right-handedness and the organization of the brain. In a study from 1977, the speech centers of 95% of the right-handed people investigated were in the left hemisphere. The rest were right-dominant for speech. The hemisphere dominant for speech in 70% of the left-handed people was also the left one. Approximately 15% of the left-handed people studied had their speech centers in the right hemisphere; the other 15% seemed to have bilateral language control. Newer studies reveal that right hemispheric speech is possibly even more rare than assumed. It occurs mostly when the left hemisphere has been injured early in life.

(10)

59

In other cases, the speech is most likely bilaterally represented in the brain. Right hemisphere language dominance is not uncommon in adults who received injury to the left hemisphere early in life. Nevertheless, the adaptability or plasticity of the nervous system decreases with age, and when left hemisphere injury occurs after puberty, chances are good that a permanent aphasia will result. Although speech and language function can be taken over by the right hemisphere if necessary, there is evidence that the right hemisphere doesn't have the same potential for speech and language specialization as the left. Maureen Dennis and Harry Whitaker monitored three children in whom one hemisphere was surgically removed during infancy to stop seizures. Two had only the right hemisphere and one only the left. At the age of ten these children were given psychological and psycholinguistic tests. Their intelligence was found to be comparable. All three could produce and comprehend normal speech. However, when tested for different sentence structures, only the child with the left hemisphere was able to maintain proficient performance when presented with complex syntactic constructions such as those in (i). Only the child with the left hemisphere judged sentences a. and b. as grammatically wrong and c. as correct: (i) a. I paid the money by the man. b. I was paid the money to the lady. c. I was paid the money by the boy.

This study suggests that the plasticity of the brain is limited in scope and that the functional asymmetry between the two hemispheres may already be effective in early childhood. (11) References: Springer, Sally & Georg Deutsch, 1993. Linkes Rechtes Gehirn. Third Edition. Heidelberg: Spektrum Akademischer Verlag. Whitaker, Haiganoosh, 1976. A Case of the Isolation of the Language Function. In: Witaker, H. & H.A.Witaker. Studies in Neurolinguistics. Vol. 1. New York: Academic Press, 1-58. Language Acquisition First language acquisition takes place very rapidly. By the time a child enters elementary school, he or she has learned much of the grammar of his or her language. The speed of acquisition and the fact that it occurs without overt instruction for all children, regardless of significant differences in their social and cultural environments, have led linguists to believe that there is some 'innate' predisposition in the human child to acquire language: The human brain is 'prewired' for language in the sense that when children are exposed to speech, certain general capacities for discovering the structures of language automatically begin to operate. This is called the 'language-faculty'. The child uses his language faculty to discover the patterns in the utterances heard around him/her. From the 'primary linguistic data' he sets up hypotheses about the grammar of the language. These hypotheses are used to produce sentences that after a period of trial and error begin to conform to the speech of adults. By itself, however, this faculty is not enough. A child growing up requires interactions with other language-users in the first two or three years of life in order to

8 (1)

60

set the 'language-faculty' into operation with a particular language, such as English. The child must also be physically capable of sending and receiving (sound) signals in a language. All infants make 'cooing' and 'babbling' noises during their first few months of life, but infants deaf at birth stop after six months. So in order to speak a language, a child must be able to hear that language being used. By itself, however, hearing language sounds is not enough. Example: In one reported case, deaf parents made sure that their son who could hear normally had ample exposure to TV and radio programs. But the child did not acquire an ability to speak or understand spoken English. What the child did learn very effectively by the age of three, however, was the use of American Sign Language - the language he used to interact with his parents. The crucial requirement for learning a language appears to be the opportunity to interact with others via language. (2) The acquisition schedule. All normal children, regardless of culture, develop language at roughly the same time, along a very similar schedule. Since the same is true for sitting up, standing, walking, using the hands and many other physical activities, it has been suggested that the language acquisition schedule follows a similar basis as the biologically determined development of motor skills. This biological schedule is tied to the maturation of the infant's brain and its lateralization process. But the biological program underlying language acquisition is dependent on an interplay of many social factors in the environment as well. The acquisition process requires a constant input of 'primary linguistic data' which forms the basis for discovering the regularities in the particular language the child is learning. The child is actively involved in acquiring his language in that he must 'figure out' the regularities in the linguistic data he hears. What the child actually acquires is a grammar - a mental system that allows one to speak and understand a language. Language develops by stages. Some stages may overlap for a short period, but the transition between them is often sudden. Whereas the stages are probably universal, there is substantial variation among children in terms of the age at which a particular stage of linguistic development occurs. 1. Pre-language stages. At about 6 months children begin to develop the articulatory movements needed to produce the speech sounds of their language. This the babbling stage. 2. The one-word or holophrastic stage. Between 12-18 months children begin to use the same string of sounds repeatedly to mean the same thing. They have learned that sounds are related to meanings and produce their first words. The single words name everyday objects: a. [da] dog b. [sa] sock c. [aj] light d. [sae:] what's that? One single form like dada functions as a phrase or sentence, perhaps meaning 'Daddy just entered the room' . This use is termed holophrastic. Further examples are, down 'the child wants down', door 'the father closed the door' etc. 3. The two-word stage. At about two years combinations like the following occur: a. baby chair c. Mommy sock b. allgone ball d. byebye boat Their interpretation is closely connected to the context: baby chair can mean a)

(3)

61

possession (= baby's chair), b) request (= put the baby in the chair) or c) a statement (= the baby is in the chair). Inflectional affixes are absent. 4. Telegraphic speech. Between 2 and 3 years the child begins producing a large number of multiple-word utterances: a. Andrew want ball b. cat drink milk c. this shoe all wet d. what her name? Because these utterances lack bound morphemes and most nonlexical categories, they resemble the style of language found in telegrams. The child has clearly developed some sentence-building capacity and can order the forms correctly. A number of grammatical inflections begin to appear as well as some simple prepositions like in, on. By 3 years the vocabulary includes hundreds of words and pronunciation is close to the adult language. (4) Phonological development. By the time children reach their first birthday (i.e. enter the one word stage) they have learned that sounds are used to express differences in meaning but their own ability to produce the sounds of their language is very limited. Certain general trends in their phonological development can be discerned: Consonant sounds are more likely to be first used correctly at the beginning of a word; final consonants emerge later (two exceptions are the early use of final [f] and [s] in English). One study showed that during the second year, [p], [b], [k], [n], [f], [d], [g], [m] and [h] are often used word-initially, but only the first five of these sounds are common word-finally. Furthermore at least eight vowels were used by the end of the second year: [I], [i], [a], [U], [Z], []], [Y] and [aI]. By age 4 all the vowels and diphthongs were in place and most consonants. The only consonants missing were [>], [ð], [o] and [¥] (and certain uses of [l], [õ], [t] and [z]). Experiment 1: The 'fis' phenomenon. There is often a difference between what children understand and what they produce. One child had a plastic inflated fish which he called 'fis'. An adult imitated the child and asked him 'Is this your fis?' The child answered 'No, my fis.' and continued to reject the adult's pronunciation until the adult said 'That is your fish.' The child then answered 'Yes, my fis.' Phonemic differences aren't recognizable by children under about 18 months. Experiment 2: Children were given two toy animals named bok and pok and asked Show me pok. Under 18 months they couldn't distinguish between the two words. Morphological development. The acquisition of bound morphemes and nonlexical categories (determiners, auxiliaries) in English takes place between the age of 20 to 36 months in a sequence with relatively little variation from child to child. The typical developmental sequence for English-speaking children is as follows: a. -ing running b. plural -s balls c. possessive -'s kitty's d. determiners the, a e. past tense -ed talked f. third person singular -s drinks g. auxiliary be This developmental sequence seems to be unrelated to the frequency of the different morphemes in the speech heard by children. The determiners the and a are

(5)

62

the most frequent grammatical morphemes. The typical relative frequency of morphemes in parental speech is: a. the, a b. -ing c. plural -s d. auxiliary be e. possessive -'s f. past tense -ed g. third person singular -s (6) The acquisition of morphological forms is often accompanied by a process of overgeneralization. The child overgeneralizes the rule of adding -s to form plurals and will say foots and mans instead of feet and men. At the same time this overgeneralization is taking place, some children also begin using irregular plurals such as men appropriately for a while, but then try the general rule on the forms producing expressions like some mens and two feets. The same is true for past tense: 1) Children may say brought or broke before they begin to use the incorrect forms bringed and breaked. At this early stage they don't use regular past tense forms like kissed, walked, or helped, which means that they probably don't know that brought or broke are past tense forms. 2) When they begin to say played, hugged and helped as well as play, hug, and help, they have figured out how to form a past tense; i.e. they have constructed the rule of past tense formation. At that point they form all past tenses by this rule. This results in overgeneralization and they no longer say brought but bringed. The acquisition of the rule overrides previously learned words. 3) Later they will learn that there are irregular forms that are exceptions to the rule and then will once more say brought. By the age of 4 years the child uses the correct forms of regular and irregular verbs. Semantic Development. By the age of 18 months the average child has a vocabulary of fifty words or more. Then the vocabulary starts to grow rapidly, sometimes by as many as ten to twelve words a day. By the ages of 2 children's vocabulary exceeds 200 words. Noun-like words are predominant in the child's early vocabulary, with verb- and adjective-like words being the next most frequent. By the age 6 most children have acquired about 5,000 different words. A major factor in lexical development is the child's ability to use contextual clues to draw inferences about the category and meaning of new words. From around 17 months, children can use the presence or absence of determiners to distinguish between proper nouns and common nouns. Experiment 1: Two-year-old children who are told that a new doll is a dax will identify a similar doll as a dax as well. However, if they are told that the new doll is Dax, they will restrict use of the new word to the doll they have actually been shown, as if it were a proper noun (the doll's name). Typical semantic errors involve overextension and underextension. Children overextend a word to include a set of perceptually similar objects. Example: dog is frequently overextended to include horses, cows and other four-legged animals. Moon can be used for a large set of round objects. Children overextend more in their production than in their comprehension. Experiment 2: A child who overextends the word dog in his own speech may point only to dogs when asked by an adult to find the dogs in a picture shown to him.

(7)

63

This suggests that children deliberately overextend words in production to compensate for their limited vocabulary. Children can also use lexical items in an overly restrictive fashion (= underextension). Example: The word dog might be used for collies, spaniels, and beagles but not for chihuahuas. This points to the tendency to focus on prototypical members of a category. (8) Syntactic Development. Radford (1990) distinguishes three stages in the syntactic development of children: the categorial stage (12-18 months), the lexical-thematic stage (18-24 months) and the function stage (from 24 months on). All children learning English go through these stages in this order although their actual ages may differ +/- 20%. During the categorial stage children learn that words belong to syntactic categories (noun, verb, preposition, adjective) and begin assigning a category to the words they know and learn. In the lexical-thematic stage the categorized words from the child's lexical are used in lexical phrases (noun phrases, verb phrases, prepositional phrases and adjective phrases). He believes that the functional categories of the grammar are still missing at this stage and the lexical phrases are characterized only by the thematic relationship they hold to their co-constituents. (1) a. cup tea, bottle juice, mummy car, Hayley dress b. open box, horse tumble, baby drive truck c. in there, out cot, in dey hole, to nice lady, on carpet d. big heavy book, good girl, dirty diaper, nice yellow pen e. man no go in there, Kathryn no like celery

Then rather suddenly from 23 to 24 months functional categories mature. The following well-formed expressions are taken from children in the functional stage at 26 months: (2) a. a bottle, the freezer, the ten little ducks, my cup of tea b. I don't want it, He can't find them, She'll have to lie on my lap. c. Have I got it? What are you saying? Is any more in there? Shall I go fetch it?

Radford's viewpoint has unleashed a debate in the linguistic literature on first language acquisition on whether children's grammar matures in the way proposed by him via the three stages discussed above or whether children know and are able to make use of functional categories from the beginning of their linguistic experience, cf. Deprez & Pierce (1997). This debate has been termed maturation vs. continuity. (9) How the acquisition process proceeds. As the linguistic ability of the child increases, it is often assumed that the child is being 'taught' the language. This view underestimates the child's role in learning a language. For the vast majority of children, no one provides any instruction on how to speak the language. A more realistic view is that the child is actively constructing a grammar. 1. Imitation. It is not possible that the child acquires his language through a process of consistently imitating adult speech alone. a. Overgeneralizations reflect the attempt to construct grammatical rules, not the

64

imitation of adult speech. b. Children are typically unable to imitate structures that they have not yet learned on their own. Example 1: P: What have you seen? C: What you have seen? Example 2: P: The sun isn't shining. C: No the sun shining. Many utterance types produced by children don't occur in or even resemble structures found in adult speech. Children must process the speech they hear in terms of their current grammatical system. 2. Correction. Nor can adult correction account for the acquisition of a grammar. a. Instead of correcting children's speech, parents usually react to the meaning of children's utterances. Example.: C: Mama isn't a boy, he's a girl. P: That's right. b. Even when adults do attempt to correct grammatical errors, their efforts have little effect. Children continue to use their own forms. Example 1: C: Nobody don't like me. P: No, say "Nobody likes me." C: Nobody don't like me. [This exchange is repeated eight times] P: No, now listen carefully; say "Nobody likes me." C: Oh! Nobody don't likes me. Example 2: C: Want other one spoon, Daddy. P: You mean, you want the other spoon. C: Yes, I want other one spoon, please Daddy. P: Can you say "the other spoon"? C: Other ... one ... spoon. P: Say "other". C: Other P: "Spoon". C: Spoon P: "Other spoon." C: Other ... spoon. Now give me other one spoon? 3. The role of Motherese. Speech addressed to young children has special properties that could help increase its understanding. Motherese consists of clear articulation with pauses between phrases and exaggerated intonation contours to signal questions, imperatives and statements. It concentrates on the present situation and contains primarily statements relating to the child's current surroundings, activities and needs. Example: P: That's right, pick up the blocks. (said as the child is picking up a box of building blocks). P: That's a puppy. (said as the child is looking at a young dog). It isn't clear whether there are real correlations between caregiver speech and child speech. For example: The relative frequency of bound morphemes in parental speech apparently has no direct effect on their order of acquisition. (10) Independence of the language faculty. There is reason to think that language

65

acquisition and other types of cognitive development are to a very significant degree independent of each other. a. Rick is a severely retarded 15-year-old whose performance on a variety of nonlinguistic tasks suggests that his general cognitive level is that of a preschool child. However, his speech contains the appropriate use of affixes, nonlexical categories and correct word order. Examples: a. She must've got me up and thrown me out of bed. b. If they get in trouble, they'd have a pillow fight. c. She's the one that walks back and forth to school. d. I wanna hear one more just for a change. Certain aspects of language (morphology and syntax) are independent of nonlinguistic types of cognitive development. This implies that the mental mechanisms responsible for the acquisition of these parts of the grammar are relatively autonomous. b. Case of severe child neglect. Genie was kept tied to a chair in a small room with virtually no opportunity to hear human speech from the age of 1 1/2 to 13 years. After long therapy and care, Genie's nonlinguistic cognitive functioning was described as relatively normal and her lexical and semantic abilities are good. Her syntax and morphology, however, are not appropriate. Examples: a. Applesauce buy store. b. Man motorcycle have. c. Want go ride Miss F. car. d. Genie have full stomach. e. Mama have baby grow up. (11) Is there a 'critical age' after which first language acquisition is no longer possible? Report in TIME on the human brain: When a baby is born, it can see and hear and smell and respond to touch, but only dimly. The brain stem, a primitive region that controls vital functions like heartbeat and breathing, has completed its wiring. Elsewhere the connections between neurons are wispy and weak. But over the first few months of life, the brain's higher centers explode with new synapses. And as dendrites and axons swell with buds and branches ... metabolism soars. By the age of two, a child's brain contains twice as many synapses and consumes twice as much energy as the brain of a normal adult. University of Chicago pediatric neurologist Dr. Peter Huttenlocher has chronicled this extraordinary epoch in brain development by autopsying the brains of infants and young children who have died unexpectedly. The number of synapses in one layer of the visual cortex, Huttenlocher reports, rises from around 2,500 per neuron at birth to as many as 18,000 about six months later. Other regions of the cortex score similarly spectacular increases but on slightly different schedules. And while these microscopic connections between nerve fibers continue to form through life, they reach their highest average densities (15,000 synapses per neuron) at around the age of two and remain at that level until the age of 10 or 11. The brain's greatest growth spurt, neuroscientists have now confirmed, draws to a close around the age of 10, when the balance between synapse creation and atrophy abruptly shifts. Over the next several years, the brain will ruthlessly destroy its weakest synapses, preserving only those that have been magically transformed by experience. This magic, once again, seems to be encoded in the genes. The short burts of electricity that travel through the brain, ..., ensure the survival of synapses by stimulating genes that promote the release of powerful growth factors and suppressing genes that encode for synapse-destroying enzymes.

66

There appears to be a series of windows for developing language. The window for acquiring syntax may close as early as five or six years of age, while the window for adding new words may never close. The ability to learn a second language is highest between birth and the age of six, then undergoes a steady and inexorable decline. Many adults still manage to learn new languages, but usually only after great struggle. (11) References: Akmajian, A., R.Demers, A.Farmer & R.Harnish, 1995. Linguistics. An Introduction to Language and Communication. Fourth Edition. Cambridge, Mass.: MIT Press. Fromkin, Victoria & R.Rodman, 1998. An Introduction to Language. Sixth Edition. New York: Harcourt Brace College Publishers. O'Grady, William, M.Dobrovolsky & Fr.Katamba, 1996. Contemporary Linguistics. An Introduction. Third Edition. New York: St. Martin's Press. Crystal, David, 1995. The Cambridge Encyclopedia of Language. Cambridge University Press. Deprez, Vivian & Amy Pierce, 1993. Negation and functional Projections in Early Grammar. Linguistic Inquiry 1, 25-67. Radford, Andrew, 1990. Syntactic Theory and the Acquisition of English Syntax. Oxford: Basil Blackwell.



Hasonló témájú dokumentumok

A csoport

- 2008-09-19 10:18:03

szolgálati szerződés fajtái

- 2008-05-22 18:15:15

10.gyakorlat

- 2009-10-07 13:24:52

Guzsik-jegyzet szakrális építészetről

- 2008-04-15 22:29:04

előadás jegyzet 3

- 2008-05-22 18:10:57

Szociológia.III

- 2010-01-09 15:08:45

Általános pszichológia jegyzet (Atkinson)

- 2008-02-18 02:16:21

A mások által feltöltött dokumentumokat értékelheted. Ha úgy ítéled meg, hogy a vizsgára való felkészülés szempontjából hasznos volt egy dokumentum, akkor adj rá sokcsillagos értékelést.
Ha hibákat tartalmaz, vagy egyéb probléma van vele, akkor keveset.
A dokumentumok sorrendje az értékelések alapján adódik. Ami fentebb van a listában, azt hasznosabbnak ítélték társaid. Az új dokumentumok pedig (értékelések hiányában) szintén a lista tetején kezdenek.

Hozzászólások

Ha észrevételed van egy dokumentummal kapcsolatban (például hibát találtál benne), akkor a Hozzászólások részben jelezheted. Az olyan jellegű kérdéseket mint pl.: A 2. feladat 4. sorából milyen átalakítással jutottunk az 5. sorban szereplő képlethez? - szintén ide érdemes írni

Egy tipp az oldalhoz! - Küldj üzenetet a szakod vagy évfolyamod összes hallgatója számára. Hasznos lehet ha választ keresel egy kérdésre, vagy mindenkivel tudatni akarsz egy információt. Ehhez használd az Üzeneteken belül a baloldali dobozban az Üzenet írását.

... 03.04/1 16 18. század 2008/2009-1 3 eloadas 5.előadás 8. előadás a csoport alkotmánytörténet államháztartás altér analízis dolgozat anyagismeret 2 economic policy elemzés etika evolúció feladatok filozófiatörténet földrajz gazdaságtörténet gyakorlat hla jogelmélet jogképesség kisebbség kriszti kultur logisztika matek jegyzet munkaerőpiac művészettörténet nemzetközi marketing objektum oktatói kiadott anyag orvosi kémia prax sejtbiosz speech sounds stilisztika szentmiklóssy természetvédelmi mérnök tűzvédelem üzleti terv vállalat gazdaságtan vállalat helye víz vizsgazh zh feladat