Phonetics 3: Articulatory phonetics: IPA and phonology

Varieties of English

• American and English pronunciation of over 75.000 words: J.C. Well’s Longman Pronunciation Dictionary

• Cover: summary in table format pf commonalities and differences

Phonology

• Phonology is the study of the sound system of a language

• While phonetics concerns the physical production, acoustic transmission, and perception of the sounds of speech

• Phonology describes the wax sounds function within a given language or across languages to encode meaning

• In other words, phonetics belong to descriptive linguistics and phonology to theoretical linguistics

Phonological analysis:

1. Which sounds does a language use?

   1. Phoneme inventory

2. What are the allowable combinations of sounds in a language?

   1. Phonotactics

3. How do sounds change depending on what is around them?

   1. Contextual changes (coarticulation, assimilation)

Phoneme inventory

• A phoneme is the smallest phonetic unit in a language that is capable of conveying a distinction in meaning, as the m of mat and the b of cat in English

• In other words, two (or more) words that differ by a single sound in the same position and that have different meanings

• Sum - sun; kill - gill; messy - meshy; feel - fill- fell

  • Minimal pair - the sounds in a minimal pair…

    • Contrast

    • Are unpredictable (they must be learned)

    • Belong to different phonemes

• In very language, certain sounds are considered to be the same sound, even though they may be phonetically distinct

• E.g. English pin (aspirated) versus spin (unaspirated)

  • Despite differences, to a native listener both words have a /p/ in them

  • When the /s/ is removed from spin, we hear bin (and not pin)

  • When an /s/ is spliced onto pin it sounds unnatural

Types of Transcription

• // Broad phonetic transcription: representation of the basic sound unity used to pronounce words

• [] narrow phonetic transcription:

  • Phonetics: representation of the phonetic detail of a sound, necessary for understanding crucial features of sounds, below the level of contrats

  • Phonology: all phonetic detail necessary for contrast

Phoneme vs. Allophone

• Phoneme: A minimal unit of sound that serves to distinguish meaning between words

• Allophone: the different phonetic realisations of a phoneme (i.e., variations in pronunciation of the same phoneme)

2 Types of distribution - type 1

1. Contrastive distribution

• When sounds can occur in the exact same phonetic environment (thereby forming  a minimal pair)

• E.g. sue - zoo, bussing - buzzing, close - clothes

  • Distinctive features

    • In English, a difference in voicing makes 2 sounds “different sounds” (e.g. pill (voiceless) and bill (voiced))

    • Many other minimal pairs of English words where the only difference is whether or not one sound is voiced (e.g. rip /rib)

→ therefore, voicing is a distinctive feature in English

2 Types of distribution - type 2

2. Complementary distribution / non-distinctive features

• When two (or more) phonetically similar sounds never occur in exactly the same environment, but in complementary or mutually-exclusive environments

  • E.g. spat - pat; spool - pool, speak - peak

• Sounds in complementary distribution

  • Are allophones of a single phoneme

  • Do not occur in minimal pairs

  • Are non-contrastive

Phonemes vs. Allophones - aspiration

• In English the p in pin is aspirated but the p in spin is not

• But, there is no minimal pairs of English words that differ only in whether or not one sound is aspirated

  • So aspiration is a non-distinctive feature in English

  • 2 sounds that differ only in aspiration are allophones of the same phoneme

  • Allophones = different versions of the “same sound”

• Koren is opposite of English (aspiration is phonemic, voicing is allophonic)

Clark Kent / Superman analogy

A comparison used to illustrate the relationship between phonemes and their allophones, where Clark Kent represents the phoneme and Superman represents its allophones, showcasing how the same underlying entity can manifest in different forms.

Phoneme inventory exercise:

• Minimal pair test can sho that 2 sounds constitute two separate phonemes in a language

• Find 2 words that differ in one sound and also differ in meaning → must be a phoneme in the language

  

  
  

• Homophones: they sound the same even though they are completely different words

  • Ate - eight, earn - urn, see - sea

Speech Variation

• We heave seen so far how sounds can be classified (ans what their acoustic properties are)

• BUT in real speech things are more complicated … there is a lack of invariance!

• Reliable constant relations between a phoneme of a language and its acoustic manifestation in speech are difficult to find

• The acoustic form varies a great deal

• Intrinsic & extrinsic distortions interact in everyday acoustic scenes

  • E.g. multi-speaker babble noise, poor acoustics, reverberation, unfamiliar (very fast) accent

Causes for variation

• Large within-talker and between-talker variability

• Coarticulation and phonologically-rules variation

• Speaking rate differences

• Noise can obscure acoustic cues

• Speech is continuous

• But: high speed of processing with average of 15 sounds per second

• Amazing ability of listener to “normalise” across exemplars

Talker variability

• When the same talker produces the same phonetic unit in the same context, it can vary (e.g. due to mood, loudness, position in a list)

• When different talkers produce the same phonetic unit, such as a simple vowel, the acoustic results vary widely

• This is because of the variability on vocal tract size and shape, and is especially different for men, women and children

• Physiological differences: size, gender, age

• Dialectal and idiolectal differences: variety characteristic of a group of speakers; variety unique to a person

Coarticulation

• The acoustic realisation of a phoneme is affected by surrounding sounds

• E.g. the place of articulation for /k/ is somewhat different in /ki/ and /ku/

Phonologically-rules variation

• Deletion = omission of a segment

• Epenthesis = successive segments are separated by an intervening segment

• Assimilation = one sound becomes more like a nearby sound; this can occur either within a word or between words

Speaking rate

• Slower speech rates typically imply more canonical pronunciation 

• faster/ normal speech rate results in many reduced speech forms (but also slow speech can be reduced)

Noisy speech

• Only rarely do we hear speech without any surrounding niceses

• Speech sounds and other sounds arrive simultaneously at our ears

• Listeners are remarkably good at understanding speech in noise (at least in their mother tongue); cocktail party effect

• Computer are still not good at it

Distorted speech

• Not only noise distorts the speech signal, other distortions are possible too (e.g. hearing loss, radio signal, …)

• Sine-wave speech is a form of artificially degraded speech

• Remez and colleagues demonstrated a dramatic change in the way in which sine-wave speech sentences are perceived, depending on listener’s specific prior knowledge

Continuous speech

• In contrast to written language, in spoken language there are no breaks between words

• Difficult to find word onsets and the speech signal matches multiple interpretation (temporarily)

Problem with acoustic articulation

• no reliable relation between a phoneme of a language and its acoustic manifestation in speech

• The acoustic form varies a great deal

  • Large within-talker and between-talker variability

  • Coarticulation

  • Speaking rate differences

  • Noise can obscure acoustic cues

  • Speech is continuous