Introduction to Contemporary Linguistics November 11, 1998 Phonetics: the physical sounds of language [WARNING! ALL OF THE PHONETIC FONTS HAVE BEEN RUINED! FOR CORRECT DETAILS, PLEASE SEE A REAL HANDOUT!!] OVERVIEW: 1. Writing phonetics 2. Articulatory phonetics 3. Acoustic phonetics 4. Speech perception 5. Video: Normal Speech Articulation ============================================================= 1. Writing phonetics >Communication means that mental things in one brain (thoughts, semantics, syntax, morphology) are transfered through the physical world (by sound or light) to another brain. [BOARD] >Phonetics is the study of this physical part of language, specifically the study of speech sounds ("phone" = "sound"). >However, this physical part of language is very hard for ordinary people to "hear": our brains are built to translate phonetics automatically into the mental structures of phonology (we'll discuss that next week). [OVERHEAD CARTOONS] >Thus ordinary writing systems, like English spelling, don't really represent phonetics very well: >For example, the letters a, e, h, i, n, r, s, t represent different sounds in "hate", "hat", "the", "thin", "heat", "rise", .... >The symbols used in Pinyin and in 注音符號 are also not true phonetic symbols. For example, both systems represent the vowels below with identical symbols. However, as your ears will tell you, the vowels are NOT phonetically the same! 注音符號 Pinyin 談 ㄊㄢˊ tan 言 ㄧㄢˊ yan >To describe real phonetics, linguists invented the International Phonetic Alphabet (IPA): a set of symbols to write all speech sounds found in all languages of the world. >The IPA represents the real phonetic sounds of words: English spelling: hate hat the thin heat rise IPA: [hejt] [haet] [D@] [TIn] [hi:t] [rajz] >The KK (Kenyon and Knott) system used to teach English in Taiwan is based on the IPA, but there are some differences: KK IPA Comment >r ... IPA [r] represents a "trill" as in Spanish; [.] is the kind of "r" that English really has >p,t,k ph,th,kh [h] = aspiration(see below) >KK is not a truly phonetic writing system, since it also shows phonological (nonphysical, mental) patterns. (see next week) >The IPA can also be used for Mandarin. 談 [than35] 言 [iEn35] >Each IPA symbol represents one segment of sound: >Some English letters actually represent two segments: jam [dZaem] >Some symbols in KK really represent two segments: KK IPA day [de] [dej] >Several symbols in 注音符號 really represent two segments: ㄤ [aN] ㄠ [au] ... ㄟ [ei] ㄡ [ou] >But it's sometimes hard to decide how many segments there are: Number of segments IPA notation 來 3 [lai35] 下 3 or 2? [Cia51] or [Ca51] 魔 2 or 3? [mWO35] or [muO35] >Segments are not the only part of phonetics. There are also phonetic properties "above" (bigger than) segments, called suprasegmentals: tone, stress, vowel length, and intonation. >Stress （重音） is marked in IPA basically as in KK, and vowel length is represented with [:], as in [a:] for along [a]. >Lexical tone is pitch used to distinguish words. IPA marks tone on a five-pitch scale: 5=highest, 1=lowest. >Notice that IPA tone marks are different from those used in 注音符號 and Pinyin: IPA Pinyin 注音符號 55 - 35 ' ˊ 213 `' ˇ 51 ` ˋ >IPA marks phonetics, not phonology. Thus Tone 3 does NOT always have the phonetic form of [213],and "Tone 0" (輕聲) can have different phonetic forms. (Such changes are due to phonology; see next week!) >Intonation is the changing pitch of a whole sentence. The way it interacts with syntax and pragmatics is complex! (1) a. Do you like to eat candy? [rising intonation] b. What do you like to eat? [falling intonation] [Does Mandarin have intonation????] >In order to get a feeling for the IPA, here is a handout that compares the IPA with 注音符號 and Pinyin, and compares the phonetic segments in Mandarin and English. [HANDOUT]. 2. Articulatory phonetics >Articulatory phonetics describes speech articulation: how the movement of the speech organs create speech sounds. >Like a car, the vocal tract is a complex machine, and every part and every action has a name. [MODEL/OVERHEAD] >Lungs. Airflow used in speech usually comes from the lungs, of course. >However, sometimes airflow is produced other ways: >Ejectives: air is "popped" out of the mouth: >Implosives: air is "sucked" into the mouth: >Clicks: air is "clicked" by the tongue or lips >Larynx. The "voice box" contains vocal folds which can vibrate at different frequencies. The glottis is the opening between the vocal folds. Tone is produced by the larynx. >Nose: Nasal sounds are produced by lowering the velum to allow airflow through the nose. The inside shape of the nose can't be changed, though. >Mouth: Lips and tongue. When these move, they cause great changes in the oral cavity, which causes great changes in the sound that comes out. >Speech sounds can be described with phonetic properties (just as word meaning was described by semantic properties last week). >Consonants have three kinds of phonetic properties: >(1) Laryngeal properties: properties of the larynx. >Voicing: >Voiced = vocal cords vibrate: [b], [d], [g] >Voiceless = no vibration: [p], [t], [k] >Aspiration: >Aspirated = noisy puff of air: [ph], [th], [kh] >Unaspirated = no puff: [p], [t], [k] >IMPORTANT NOTE: >Mandarin only uses aspiration, not voicing: ㄅ [p] ㄆ [ph] ㄉ [t] ㄊ [th] ㄍ [k] ㄎ [kh] >English uses voicing and aspiration together: "b" [b] "p" [ph] "d" [d] "t" [th] "g" [g] "k" [kh] >Other languages like French only use voicing: e.g. "b" [b] "p" [p] >Still other languages, like Taiwanese, use voicing and aspiration separately: e.g. [bi53] [pi53] [phi53] 米 比 粑／疤 >(2) Place of articulation: where the sound is made. >Labial: uses the lips >Bilabial: both lips: [b], [p], [m] >Labio-dental: lips + teeth: [f], [v] >Coronal: uses the tongue blade or tongue tip >Alveolar: tongue tip on alveolar ridge: [t], [d] >Dental: tongue tip on teeth. >Palatal: tongue blade near palate: >Alveolo-palatal: tongue blade between palate and alveolar ridge: >Retroflex: tongue tip curled back: >Velar: tongue body near velum: [k], [g], [x] >Glottal: uses the larynx only: [h], [?] >Glottal stop [?]: as in 晚安 [uan21?an55] >(3) Manner of articulation: how the sound is made. >Stops: airflow through mouth is stopped >Nasals: airflow only through nose:[m], [n], [N] >Plosives: airflow through mouth: [p], [b], ... >Continuants: airflow through mouth not stopped >Fricatives: airflow has friction: [f],[v], [s], ... >Liquids: [l], [r], ... >Glides: [j], [w] ([h]?) >Affricates: stop + fricative: [ts], ... >NOTE: sonorants = nasals, liquids, glides, vowels (they have a "smooth", less "noisy" sound) obstruents = the rest (plosives, fricatives, affricates) >Vowels have four kinds of phonetic properties: >(1) Height: how the high tongue body is: >High vowels: [i], [u], [y] >Mid vowels: [e], [o], [@] >Low vowels: [a], [ae] >(2) Backness: how far back the tongue body is: >Back vowels: [u], [o], [a] >Central vowels: [@] >Front vowels: [i], [e], [y] >(3) Roundness: whether or not the lips are rounded: >Rounded vowels: [u], [o], [y] >Unrounded vowels: [a], [i], [e] >(4) Tenseness: how far the tongue body is from the "lazy center" of the mouth: >Tense vowels: [i], [u], [e], [o] >Lax vowels: [I], [U], [E], [O], [@] >The laxest, "laziest" vowel is [@]: schwa. >By the way, human beings can only produce so many vowels because our tongue is "bent". Our ancient pre- human ancestors had a "flat" tongue and so couldn't produce as many vowels [OVERHEAD]. >In real speech, the articulators are all moving around at the same time, very quickly. Thus even IPA segments are not truly phonetic, since real speech shows coartication: articulations that overlap in time (e.g. the tongue moves to make a vowel while the lips are still producing a consonant). 3. Acoustic phonetics >Acoustic phonetics describes speech acoustics: the physical sounds produced by the vocal organs. >Example: the acoustics of laryngeal properties: [OVERHEAD] >Voicing appears as regular beats in the sound wave. >Higher tones have higher frequency = beats are more frequent in every unit of time. >Voiced consonants (e.g. [b]) have a short Voice Onset Time (VOT): voicing starts early, often before the mouth opens. >Voiceless aspirated consonants (e.g. [ph]) have a very long VOT: voicing starts very late. >Voiceless unaspirated consonants (e.g. [p]) have a VOT of a medium duration. >Spectrogram: a visual display of speech acoustics: [OVERHEADS] >Time: left to right. >Frequency: up (higher = higher frequency). >Concentration of energy: darkness. >Formant: a dark band in a spectrogram showing a concentration of energy at a particular frequency. >Front, unrounded vowels (e.g. [i]) have a higher formant than back, rounded vowels (e.g. [u]). [OVERHEAD] >Place of articulation in consonants also affects formants [OVERHEAD] 4. Speech perception >As noted above, we don't hear the physical reality of speech. >Categorical perception: the illusion in which acoustically variable speech is heard as discrete, distinct sounds >For example, using a computer to create fake speech, we can vary VOT from very short to very long. When it's very short, the syllable will sound like [ba]. When it's very long, it will sound like [pha]. But in between, the sound does NOT gradually change from [ba] to [pha]. Instead, it sounds like there is a sudden jump from [ba] to [pha]. [OVERHEAD] >Listen. When does [ba] change to [pha]]? [TAPE]. 1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17-18-19-20-21 5. Video: Normal Speech Articulation. >We only have time for a short glimpse of this video. Just notice two things: >During speech, the articulators are really jumping around like crazy, especially the tongue! >There is a lot of coarticulation: different articulators are moving at the same time.