What do you think?
Rate this book
Understanding Phonetics
Assuming little or no background knowledge and using original examples and exercises (with answers supplied), Understanding Phonetics provides you with an accessible introduction to the basics of phonetics and a comprehensive analysis of traditional phonetic theory - the articulation and physical characteristics of speech sounds. Examples from a wide range of languages are presented throughout using symbols of the International Phonetic Alphabet. To help you develop your skills in this alphabet, Understanding Phonetics includes ear-training exercises that are freely available online, along with audio files of authentic listening material, for you to download from www.routledge.com/cw/ashby. Understanding Phonetics outlines the production of consonants, vowels, phonation types, pitch and intonation, and aspects of connected speech. Reading through chapter by chapter, you will see your knowledge develop as you engage in the step-by-step phonetic study of a selected word. Understanding Phonetics is designed to be used not only as a class textbook but also for self-study. It can be read systematically or used for reference purposes.
- GenresLinguistics
244 pages, Paperback
First published January 1, 2011
2 people are currently reading
17 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 of 1 review
March 29, 2024
This was my first systemic study of phonetics. (Actually a shameless lie, before was Ladefoged, 'Vowels and Consonants'. That I didn't understand his writing is entirely on me). But I really should've started with this one. It's not an easy reading either, it took me a while to stomach it, and I chewed a lot. I had many blank spots on my knowledge map, there are fewer now, at least not the same, while the map has grown bigger. But more importantly, it put the pieces together into a coherent system. Sometimes reading wiki and watching lectures is just not enough, you need a good old textbook.
The book is about phonetics, not phonology. In engineer's terms, phonetics describes the analogue signal on the wire, whereas phonology is the digital data that can be assigned meaning. The first is continuous, the second is discrete. Phonetics is also the cornerstone of phonology.
As with all good things, it comes in three forms. Articulatory phonetics studies the making of the sounds, it all starts here. Acoustic phonetics studies the raw physical data, nothing can reach the ear that is not in the waveform. But it also contains a ton of noise, full of features irrelevant to the ear. The truth is thus in the ear of the beholder which leads us to the auditory phonetics.
>> A recipe for CoNSoNaNTS
Consonants are a good starting point for the course. They are conveniently described with three parameters: Voicing, place and manner of articulation. VPM for short.
1. It all begins with the airstream and we have several types:
- Pulmonic egressive. This one is universal, and probably the only one you use.
- Pulmonic ingressive. Not found in languages.
- Glottalic egressive AKA ejectives. Hint: Ejectives can be practised with emptied lungs.
- Glottalic ingressive AKA implosives are rare. Surprisingly, these can be voices since the oral chamber is separated from the larynx. The vocal folds can work while the mouth sucks the air in.
- Oral egressive. Not found in languages.
- Oral ingressive AKA clicks are the rarest in use. They also take accompaniments, a topic covered extensively by Ladefoged.
2. To this, we can add voice (Fx - fundamental frequency) in the larynx. It can be a modal or normal voice, a creaky voice (vocal fry), or a breathy voice (as when running out of breath or whispering).
The timing of the voice is crucial. It's called voice onset time (VOT). Aspirated consonants are explained as continued with voiceless vowels. They can even be pre-aspirated, which being a peripheral feature, is not covered in detail. I'm still lost when it comes to voiced aspirates (murmured stops). It seems not to be measurable on the VOT scale, but it rather uses a different voicing technique (breathy instead of modal).
3. As the air passes through the vocal tract, the obstacles and resonators give the waveform a particular shape.
Place of Articulation means anatomy. There are a bunch of passive and active articulators, and the master of them all is the tongue. In fact, it's so flexible we can use its parts independently, so it can contribute to double articulation.
Manner of articulation is the technique. A closure produces a plosive/stop, a narrow approximation gives fricatives, a wider approximation produces approximants. Along the way, I figured out why Castilian sounds so pharyngeal to me, their velar fricative is pronounced as an approximant in quick speech.
Now the consonant is ready, we can add the topping. One can glottalise, palatalise (softening), velarise (dark l), pharyngealise (Arabic emphatics), labialise and nasalise. It's called double articulation, and the book covers them in passing.
The stream can also be channelled centrally or laterally along the tongue. One can easily feel the path when sucking in cold air.
Further variation comes from closure release techniques. Applosives (inaudible releases) can be released but be masked by a glottal stop (effectively glottalised). The opposite is also possible, [?k] will sound just like [k:]. The silent release can also be implemented through nasal release.
Syllable-final devoicing is a common pattern. In such positions, phonemic contrast may be neutralised.
>> V is for Vowels.
Vowels are notoriously hard to describe and put on the map. The clue came from Daniel Jones, he described the so-called cardinal vowels and placed them in a quadrilateral (sometimes triangular) space roughly corresponding to the tongue positions. And he was wrong. Nowadays vowels are described by their acoustic properties and auditory perception. Still, that notion of tongue positions is enshrined in the IPA terminology and it's a handy way to think of and teach them. We speak of vowel Backness/Openness/Roundedness. The author used BOR for short.
Fast forward: We have vowel quantity, length clipping, diphthongs, monophthongisation (smoothing, compression), diphthongisation (breaking), rhotic r-coloring, friction.
Note that vowel quantity is a relative measure, some long vowels pre-fortis or in a long rhythmic foot may be reduced (clipped) significantly.
Patricia did a great job explaining the spectrograms. You start with the waveform, then take a snapshot in time, decompose the complex wave into components (think FFT), and plot amplitude against frequency for each simple wave. One can already see the formants here. Now you extend this frequency domain graph by bringing back the time dimension. This 3D data is flattened by plotting frequency vs time, with the amplitude presented qualitatively with colour darkness. And voila, you have your spectrogram. Now I finally enjoy my Praat graphs. They make sense to me.
I found more on the cardinal vowels at https://www.englishspeechservices.com.... In fact, the vowel chart can easily be reinterpreted based on formants, namely F1 (top-down) vs F2 (RTL). Also F2 > F1 at all times. It's also noteworthy that the F1/F2 ratio is what matters, not the absolute values. Be that not the case, the same sound system could not be used by males, females and children alike.
===
With consonants and vowels covered, it's time to unlearn them. The sounds can be grouped differently, producing dichotomies like obstruents vs sonorants or fortis vs lenis. Different systems highlight different properties in common. Sonorant consonants share a lot with vowels, and lenis vs fortis can be more detailed than just voiced vs voiceless.
>> Connectedspeech'n'segments
Now it's time to put the sounds to work. We don't pick up the absolute values of formants, it's the degree of transition of formants (F2 for stops, F3 for l/r) that we hear and recognise as a distinct sound, and it helps a lot with place of articulation perception. Similarly, the rate of transition signifies the manner, stop, approximate or even diphthong. This is the locus theory.
When transitioning between sounds we also face interference effects. Coarticulation smoothens connected speech and is phonetic by nature. Its phonological twin is called assimilation.
>> Suprasegmentals
The final chapter gallops over stress types, accent, pitch accent, tones and intonation. And don't forget isochrony. Surely, suprasegmentals deserve a textbook on their own.
In the end, once again, phonetics operates in a continuous domain. This means there is an endless variation to the closures, articulatory approximation, VOT timing, etc. There is accordingly a ton of IPA diacritics used in narrow phonetic transcriptions. If you are in pain a parametric diagram comes to your rescue. These variations are often unnoticeable to the untrained ear, they are fluid and transient. I believe they provide a critical clue to the understanding of language change.
P.S. All English examples are given in the British Received Pronunciation.
Errata: The Georgian Mkhedruli is dated by III AD. She surely meant Asomtavruli, and it's V AD. Check your facts (╯ಠ╭╮ರ)╯ or don't use them.
The book is about phonetics, not phonology. In engineer's terms, phonetics describes the analogue signal on the wire, whereas phonology is the digital data that can be assigned meaning. The first is continuous, the second is discrete. Phonetics is also the cornerstone of phonology.
As with all good things, it comes in three forms. Articulatory phonetics studies the making of the sounds, it all starts here. Acoustic phonetics studies the raw physical data, nothing can reach the ear that is not in the waveform. But it also contains a ton of noise, full of features irrelevant to the ear. The truth is thus in the ear of the beholder which leads us to the auditory phonetics.
>> A recipe for CoNSoNaNTS
Consonants are a good starting point for the course. They are conveniently described with three parameters: Voicing, place and manner of articulation. VPM for short.
1. It all begins with the airstream and we have several types:
- Pulmonic egressive. This one is universal, and probably the only one you use.
- Pulmonic ingressive. Not found in languages.
- Glottalic egressive AKA ejectives. Hint: Ejectives can be practised with emptied lungs.
- Glottalic ingressive AKA implosives are rare. Surprisingly, these can be voices since the oral chamber is separated from the larynx. The vocal folds can work while the mouth sucks the air in.
- Oral egressive. Not found in languages.
- Oral ingressive AKA clicks are the rarest in use. They also take accompaniments, a topic covered extensively by Ladefoged.
2. To this, we can add voice (Fx - fundamental frequency) in the larynx. It can be a modal or normal voice, a creaky voice (vocal fry), or a breathy voice (as when running out of breath or whispering).
The timing of the voice is crucial. It's called voice onset time (VOT). Aspirated consonants are explained as continued with voiceless vowels. They can even be pre-aspirated, which being a peripheral feature, is not covered in detail. I'm still lost when it comes to voiced aspirates (murmured stops). It seems not to be measurable on the VOT scale, but it rather uses a different voicing technique (breathy instead of modal).
3. As the air passes through the vocal tract, the obstacles and resonators give the waveform a particular shape.
Place of Articulation means anatomy. There are a bunch of passive and active articulators, and the master of them all is the tongue. In fact, it's so flexible we can use its parts independently, so it can contribute to double articulation.
Manner of articulation is the technique. A closure produces a plosive/stop, a narrow approximation gives fricatives, a wider approximation produces approximants. Along the way, I figured out why Castilian sounds so pharyngeal to me, their velar fricative is pronounced as an approximant in quick speech.
Now the consonant is ready, we can add the topping. One can glottalise, palatalise (softening), velarise (dark l), pharyngealise (Arabic emphatics), labialise and nasalise. It's called double articulation, and the book covers them in passing.
The stream can also be channelled centrally or laterally along the tongue. One can easily feel the path when sucking in cold air.
Further variation comes from closure release techniques. Applosives (inaudible releases) can be released but be masked by a glottal stop (effectively glottalised). The opposite is also possible, [?k] will sound just like [k:]. The silent release can also be implemented through nasal release.
Syllable-final devoicing is a common pattern. In such positions, phonemic contrast may be neutralised.
>> V is for Vowels.
Vowels are notoriously hard to describe and put on the map. The clue came from Daniel Jones, he described the so-called cardinal vowels and placed them in a quadrilateral (sometimes triangular) space roughly corresponding to the tongue positions. And he was wrong. Nowadays vowels are described by their acoustic properties and auditory perception. Still, that notion of tongue positions is enshrined in the IPA terminology and it's a handy way to think of and teach them. We speak of vowel Backness/Openness/Roundedness. The author used BOR for short.
Fast forward: We have vowel quantity, length clipping, diphthongs, monophthongisation (smoothing, compression), diphthongisation (breaking), rhotic r-coloring, friction.
Note that vowel quantity is a relative measure, some long vowels pre-fortis or in a long rhythmic foot may be reduced (clipped) significantly.
Patricia did a great job explaining the spectrograms. You start with the waveform, then take a snapshot in time, decompose the complex wave into components (think FFT), and plot amplitude against frequency for each simple wave. One can already see the formants here. Now you extend this frequency domain graph by bringing back the time dimension. This 3D data is flattened by plotting frequency vs time, with the amplitude presented qualitatively with colour darkness. And voila, you have your spectrogram. Now I finally enjoy my Praat graphs. They make sense to me.
I found more on the cardinal vowels at https://www.englishspeechservices.com.... In fact, the vowel chart can easily be reinterpreted based on formants, namely F1 (top-down) vs F2 (RTL). Also F2 > F1 at all times. It's also noteworthy that the F1/F2 ratio is what matters, not the absolute values. Be that not the case, the same sound system could not be used by males, females and children alike.
===
With consonants and vowels covered, it's time to unlearn them. The sounds can be grouped differently, producing dichotomies like obstruents vs sonorants or fortis vs lenis. Different systems highlight different properties in common. Sonorant consonants share a lot with vowels, and lenis vs fortis can be more detailed than just voiced vs voiceless.
>> Connectedspeech'n'segments
Now it's time to put the sounds to work. We don't pick up the absolute values of formants, it's the degree of transition of formants (F2 for stops, F3 for l/r) that we hear and recognise as a distinct sound, and it helps a lot with place of articulation perception. Similarly, the rate of transition signifies the manner, stop, approximate or even diphthong. This is the locus theory.
When transitioning between sounds we also face interference effects. Coarticulation smoothens connected speech and is phonetic by nature. Its phonological twin is called assimilation.
>> Suprasegmentals
The final chapter gallops over stress types, accent, pitch accent, tones and intonation. And don't forget isochrony. Surely, suprasegmentals deserve a textbook on their own.
In the end, once again, phonetics operates in a continuous domain. This means there is an endless variation to the closures, articulatory approximation, VOT timing, etc. There is accordingly a ton of IPA diacritics used in narrow phonetic transcriptions. If you are in pain a parametric diagram comes to your rescue. These variations are often unnoticeable to the untrained ear, they are fluid and transient. I believe they provide a critical clue to the understanding of language change.
P.S. All English examples are given in the British Received Pronunciation.
Errata: The Georgian Mkhedruli is dated by III AD. She surely meant Asomtavruli, and it's V AD. Check your facts (╯ಠ╭╮ರ)╯ or don't use them.
Displaying 1 of 1 review