Questions & explanations
1. How does syllable weight affect stress placement in Latin?
In Latin, stress falls on the second-to-last syllable if it is heavy, otherwise on the third-to-last. For example, 'amicus' (friend) has stress on 'mi' because it is heavy (long vowel). In 'puer' (boy), the second-to-last syllable 'pu' is light, so stress falls on the first syllable 'pu'? Actually, 'puer' has two syllables: 'pu' (light) and 'er' (light? 'er' has a coda? 'r' is a consonant, so 'er' is heavy? Wait, 'er' has a short vowel + coda, so it's heavy. Then stress would be on 'er'? No, Latin stress is on the penult if heavy, else on the antepenult. For 'puer', penult is 'pu' (light), so stress on antepenult? But 'puer' only has two syllables, so stress on first? Actually, two-syllable words always stress the first? I need to be accurate. Let's use a clear example: 'civitas' (city) — 'vi' is heavy (long i), so stress on 'vi'. 'facilis' (easy) — 'ci' is light, 'fa' is light, so stress on 'fa'? Actually, 'facilis' has three syllables: 'fa' (light), 'ci' (light), 'lis' (heavy? 'lis' has short i + s, so heavy). Penult 'ci' is light, so stress on antepenult 'fa'. Yes.
2. How does the OCP apply to place features in consonant clusters?
The OCP can forbid two adjacent consonants with the same place of articulation. For example, in English, the prefix 'in-' becomes 'im-' before labial sounds (e.g., 'impossible') because the OCP disallows two labial nasals /nm/? Actually, /n/ becomes /m/ before /p/, /b/, /m/ — the OCP avoids two labial consonants? More precisely, the OCP avoids identical place features in adjacent segments. So 'in-' + 'possible' becomes 'impossible' to avoid two labials? Wait: /n/ is coronal, /p/ is labial — not identical. Better example: in some languages, a sequence like /pab/ might be simplified to /pab/... Actually, a clear example: in Arabic, the definite article 'al-' assimilates to a following coronal consonant (e.g., 'al-shams' becomes 'ash-shams') because the OCP disallows two identical coronal features? Actually, it's assimilation, but the OCP triggers it. So the OCP can cause place assimilation to avoid identical place features.
3. Why did rhotacism not affect all 's' sounds in Latin?
Rhotacism only affected 's' sounds that were between vowels. For example, in Latin 'casa' (house), the 's' was between vowels, but it did not change because it was a later development? Actually, 'casa' kept 's' because it was not in the right environment? Wait: 'casa' had 's' between vowels, but it didn't change because rhotacism happened earlier? Let me correct: In Latin, rhotacism occurred around the 4th century BC, and it affected all intervocallic 's' that were not from an earlier 'ss' or other sources. So 'casa' might have been a later borrowing? Actually, 'casa' is a later word. In general, rhotacism applied to inherited 's' between vowels, but not to 's' from other sources or later loans.
4. Compare the use of phonetic transcription in a beginner's textbook versus an advanced pronunciation course.
In a beginner's textbook, phonetic transcription is simple and used mainly for new vocabulary. It often uses a broad transcription with only basic symbols, and the focus is on individual word pronunciation. The teacher may introduce a few symbols at a time. In an advanced pronunciation course, transcription is more detailed, including narrow transcription with diacritics for aspiration, length, and tone. Students learn to transcribe whole sentences and practice connected speech features like linking and elision. They also learn to use transcription to analyze their own speech and correct errors. The advanced course assumes students already know the basic symbols and focuses on fine-tuning.
5. Transcribe the word 'gogo' in a language that uses clicks, assuming the first 'g' is a click?
If the first 'g' is a click, it might be a dental click [ǀ] or a lateral click [ǁ]. For example, in Xhosa, the word for 'walk' is 'hamba' with no click, but a word like 'gogo' (grandmother) actually has a click? Actually, 'gogo' in Zulu means grandmother and has a click? Wait, no: 'gogo' is not a click word. Let's use a correct example: in Xhosa, 'icici' (earring) has a dental click? Actually, better: the word 'qongqothwane' (click beetle) has a palatal click [ǃ]. So a transcription might be [ǃoŋǃotʰwanɛ]. But to answer simply: if the first sound is a click, write the click symbol, e.g., [ǀoǀo] for a word with two dental clicks. However, real words have vowels and other consonants.
6. How does a speech therapist use phonetic transcription to tell the difference between a lisp and a tongue thrust?
A lisp often involves the tongue coming out between the teeth for /s/ and /z/, producing a sound like /θ/ (as in 'thin'). The therapist would transcribe 'sun' as /θʌn/ instead of /sʌn/. A tongue thrust, on the other hand, may cause the tongue to push forward during speech, affecting many sounds. The transcription might show multiple sounds being produced with the tongue in a forward position, like /t/ and /d/ becoming more dental. By comparing the client's transcription to the target, the therapist can see the pattern. For example, if all alveolar sounds (like /t/, /d/, /n/) are dentalized, it suggests a tongue thrust. The transcription provides objective evidence for diagnosis.
7. How can phonetic transcription help a speech therapist decide if a child's speech error is developmental or requires intervention?
The therapist compares the child's transcription to typical developmental norms. For example, many children replace /r/ with /w/ until age 5, so a transcription of 'wabbit' for 'rabbit' in a 3-year-old is normal. But if the same error persists at age 7, it may need therapy. The transcription also shows if the error affects many sounds or just one. If the child consistently substitutes /t/ for /k/ and /d/ for /ɡ/ (a pattern called 'fronting'), it might be a phonological process that should resolve by age 3.5. If it continues, intervention is needed. The transcription provides clear data to compare with age-appropriate speech. It also helps measure progress after therapy.
8. Give an example of a phonetic transcription that shows a child's speech error, and explain what the error is.
A child says 'gog' for 'dog'. The target transcription for 'dog' is /dɒɡ/, but the child's production is /ɡɒɡ/. The transcription shows that the child replaces the initial /d/ with /ɡ/. This is called 'fronting' if /d/ is replaced by /ɡ/ (actually, /d/ is alveolar, /ɡ/ is velar, so it's backing; but a common error is replacing /k/ with /t/). More accurately, if the child says 'tar' for 'car', the target /kɑːr/ becomes /tɑːr/. The error is 'velar fronting', where sounds made at the back of the mouth (velar) are moved to the front (alveolar). The transcription /tɑːr/ clearly shows the substitution of /t/ for /k/. This helps the therapist target the specific sound.
9. What is phonetic transcription used for in forensic linguistics?
In forensic linguistics, phonetic transcription is used to analyze spoken evidence, such as recorded phone calls or interviews. Experts transcribe the speech using the International Phonetic Alphabet (IPA) to capture exact sounds, including hesitations, stress, and intonation. This helps identify speakers, detect disguised voices, or verify the authenticity of a recording. For example, a transcript might show that a speaker uses a specific regional accent. It can also reveal emotional state through pitch and rhythm. Phonetic transcription provides a written record that can be examined in court. It is a tool for comparing speech samples and finding matches.
10. What does 'perceptual weighting' mean when we listen to speech sounds?
Perceptual weighting means that listeners pay more attention to some acoustic cues than others when identifying a phoneme. For example, in English, the difference between 'b' and 'p' relies on voice onset time (VOT), which is the time between the release of air and the start of vocal cord vibration. Listeners weight this cue heavily. However, other cues like formant transitions (changes in sound frequencies) also matter. The weight given to each cue can change depending on the surrounding sounds or the listener's language experience. This shows that phoneme identification is not based on a single cue but on a combination of cues with different importance.
11. Compare the use of phonetic transcription in studying first language acquisition versus second language acquisition.
In first language acquisition, transcription focuses on how children gradually master the sounds of their native language, often showing systematic simplifications. In second language acquisition, transcription is used to see how learners pronounce sounds that are not in their first language. For example, a Japanese learner of English might have trouble with [l] and [r], and transcription reveals which sound they use instead. Both fields use similar transcription methods, but the patterns differ: children follow universal stages, while L2 learners are influenced by their first language. Transcription helps identify specific problem sounds for teaching.
12. Compare the use of unsupervised learning (e.g., autoencoders) versus supervised learning for discovering phonetic categories in an unknown language.
Supervised learning requires labeled data (e.g., hand-annotated phonemes), which is often unavailable for understudied languages. Unsupervised learning, like autoencoders, can learn latent representations of speech sounds without labels, clustering similar acoustic patterns. For example, an autoencoder trained on raw audio might learn to separate vowels from consonants based on spectral shape. However, the discovered categories may not align with phonemic categories (e.g., they might split allophones). Supervised learning, when labels are available, directly maps to linguistic categories but requires expert knowledge. A hybrid approach is often best.