Rating: 8.2/10.
Intro textbook on English phonology, starts with a general overview of the theory of phonetics and phonology, but mostly focuses on the specific patterns of consonants, vowels, stress, and intonation patterns of English, including several of its dialects. It also describes models of difficulties faced by second language speakers attempting to learn English pronunciation.
Chapter 1. Overview of English sounds by place and manner of articulation. Many voiced consonants, like “b,” are actually partially devoiced and, compared to other languages, the amount of aspiration differs significantly between languages. Vowels are continuous and the same symbol might sound somewhat different in different languages. Some symbols in the IPA do not exist in English, and some are useful for representing disordered speech. Secondary articulation includes things like lip rounding. Other factors beyond consonants or vowels include pitch, stress, and length.
Chapter 2. Sounds are different phonemes if there is a minimal pair where the same position and environment produce different sounds, often this is not possible, but a near minimal pair is where the differences are irrelevant. The opposite case is if two sounds can never occur in the same environment, then they are allophones of a phoneme in complementary distribution, and allophones must be phonetically similar, although there is some subjectivity in what counts as similar. In many cases, sounds in one language might be allophones but are phonemes in a different language. Free variation is also possible and can be in general cases of phonemes or only in specific words. Phonological analysis is useful for designing orthographies and understanding non-native language interference.
Chapter 3. Consonants. English b/d/g are voiced only between vowels and partially voiced otherwise, so sometimes this is called the fortis-lenis distinction instead of voiced versus voiceless. A voiced stop in the final position makes a preceding vowel longer than a voiceless stop, and there is often no audible release at the end. There are many assimilation processes where a stop consonant is often dropped when followed by another one. T/D can have flapping before unstressed syllables; this can also occur between words but never before a stressed syllable. T can also be replaced by a glottal stop in the final position. Affricates are also only completely voiced between vowels, and if two of the same consonants appear in consecutive words, it is one consonant that is lengthened. The suffix “s” depends on the previous sound, whether it is a sibilant, voiced, or unvoiced. Some features experience sociolinguistic variation, such as whether the “r” is dropped in the New York dialect.
Chapter 4. Vowels are classified mainly along two dimensions: high vs low, which refers to how open the mouth is, and front vs back, which is the part of the tongue is used to produce the sound. It is often useful to make a tense-lax distinction in English, lax vowels cannot occur in stressed open syllables or before an ‘r.’ Before ‘r,’ vowels get partially neutralized and r-colored. There are at least three diphthongs in English vowels, and several vowels are diphthongized to various extents depending on the situation and the dialect. Non-rhotic dialects use the centering diphthong in place of an ‘r’ final. The reduced vowel schwa occurs in unstressed syllables, but every other vowel can occur in unstressed positions and is not always reduced to schwa. Many function words have weak forms that are used most of the time, except when they are emphasized; then, the strong form is used.
Chapter 5. Speech can be analyzed in a spectrogram, and in this analysis, vowels are determined by formants. The F1 determines high versus low vowels, and the difference between F2-F1 determines front versus back. Duration is also an important indicator. If the speaker is male or female, then the pitch (F0) is higher, and all the formants are also higher in women speakers. For stop consonants, which consonant it is can be identified by the duration of the stop, the length of the previous vowel, and the formants after the stop; only occasionally is it helpful to look at the voicing. Some fricatives, like “s” and “sh,” have very different energy levels, but others, like “f” and “th,” are quite similar, and there is only a slight difference in the formants. Affricates are similar to a stop plus a fricative, but they are much faster. Glides, liquids, and nasals are mainly identified by the formants, such as lower F1 for nasals. Analysis of spectrograms is useful for treating speech disorders.
Chapter 6. Syllables can be decomposed into an onset and a rhyme, which is useful for explaining cases where consonant clusters like “bm” cannot occur within a syllable but are possible across syllables. There is some disagreement about syllable boundaries between speakers, such as with “r,” but mostly speakers agree. First, find peaks in sonority (where the vowels have the highest sonority, followed by glides, nasals, and stop consonants, which have the lowest). Then, for each sonority peak, take the maximal onset that is permissible in the language. In onsets, sonority must rise in the two consonants, with the exception of “s,” which can occur before (so there can be up to three consonants in onset clusters). In the codas, sonority generally falls, and up to triple consonant combinations are possible; some combinations are only possible in grammatical endings like “d” or “s.” Written syllabification follows different rules, such as not breaking up affixes. Sonority levels can explain the difficulty in speech disorders and for second language learners; eg: onsets with a small jump in sonority are more difficult to pronounce.
Chapter 7. Most words in English are stressed on the penult (second-to-last) syllable unless the vowel is unstressable, in which case the stress is somewhere else. However, there are exceptions, such as pairs where the noun has stress on the first syllable and the verb on the second. Secondary stress is lighter and does not have the pitch-changing tonic accent that primary stress does. Affix stress can be grouped into three general categories: those that are always stressed, those that don’t change the stress (which includes all inflectional affixes), and those that shift the stress to the preceding syllable. There are many examples of words that are stressed differently in American and British English. Intonation – tonic accent is usually on the last lexical item but can be elsewhere to emphasize a particular piece of information. There are many possibilities in rising versus falling intonation; falling intonation tends to be used for declarative sentences and WH questions, while rising intonation is used for yes-no questions or to express surprise or uncertainty; there are also combinations of falling and then rising intonation.
Chapter 8. L2 accents are caused by a phoneme that is missing in L1 (eg: most languages don’t have the “th” sound), or by having different allophone distributions (eg: the Spanish “d”), and having different phonetic properties (eg: the Spanish “r”). Many language speakers under-differentiate English vowels, have trouble with consonant clusters and stress timing. The next few sections detail typical pronunciation errors of speakers of various languages, including Spanish, Arabic, German, French, Portuguese, Persian, etc.
L2 learners usually learn a phonetic split in allophones in the native language in basic contexts before they do in derived contexts, and they may sometimes learn the split into two phonemes and then later adjust these phonemes to the correct phonetic values. Markedness is a feature that is more rare cross-linguistically and tends to be harder to acquire, a factor beyond just the absence of a feature in L1; for example, it is easier for Spanish speakers to acquire the “sm” than the “sp” consonant cluster because the latter is more marked and it is a greater violation of the cross-linguistic sonority principle.
Optimality Theory (OT) is where a ranked set of constraints, typically markedness (to avoid certain features) and faithfulness (to avoid changing the input), and output form is the one that has the least severe constraint violation. OT is useful for modeling how L2 speakers produce different words. Perception of non-native sounds also involves several models of how they are assimilated into native categories, which explain the non-uniform success of differentiation based on phonetic distance.
Chapter 9. English has a lot of irregularity in spelling due to many historical sound changes like the Great Vowel Shift and borrowing from various languages. The same letter may have multiple possible pronunciations, and the same phoneme may be represented by multiple possible orthographies. However, morphemes often retain the same spelling despite having different pronunciations, but not always, depending its origin.
