Articulatory Phonetics (1): How Speech Sounds Are Produced

Introduction

Our speech apparatus, from the lungs to the various organs in the mouth, was not originally built for the purpose of speech. Our lungs provide the cells in the body with oxygen and expel carbon dioxide produced by their cellular processes. Our larynx, what is called the voice box, was not primarily built for the purpose of producing sound; it is there to close off the passage of food into the lungs while eating (It is also closed off when we need to suppress air from the lungs in strenuous action, like when we pick up heavy items). Our mouth organs, the tongue and teeth and lips, were to help with an initial processing of food. However, “in the course of evolution, all the organs of speech have developed in very specialized ways often quite remote from their original purpose” (Collins and Mees 2013, p.29).

We are going to begin this section by taking a look at how speech is produced, the study of which falls under the purview of phonetics. We will start our journey from
the source of speech, i.e. breath, and follow the flow of air as it travels past the various organs of speech. And we will see how those organs act upon this source of energy in an extremely well-organized and integrated manner to produce the various sounds of human speech. This area of phonetics is often mislabeled articulatory phonetics, as Catford (2001) observes, since articulation is only part of the picture: articulation is essentially the action of the articulators, tongue and teeth and lips, for example. Catford proposes that we call this organic phonetics so as to rightly attribute the action of speech production to all the organs involved in the production of speech.

-----------------------------------------------------------------------------------------------------------------------

Respiration & Phonation

The diaphragm is a thin sheet of skeletal muscle that separates the abdomen from the chest. It is the primary muscle involved in respiration, the action of breathing. When you breathe in (inhale), the diaphragm contracts allowing the lungs to expand. During exhalation (breathing out), the diaphragm relaxes to squeeze the lungs and push the air out. Control over this process, especially as air escapes the lungs and travels higher up in the trachea (breathing tube), is the first crucial step in the production of speech.

View a 3D image of the above-mentioned structures here

The air traveling up the trachea passes through the larynx, the voice box, located at the top of the trachea. This organ is essential for phonation, or voicing. Inside the larynx, there are two adjustable stretches of muscle tissue called the vocal cords. These are attached to cartilages and can be spread apart leaving the glottis (area between the vocal cords) open, or brought together to narrow or close off the glottis and therefore block the passage of air through the larynx.

See a picture of the larynx and glottis here, along with a short clip demonstrating how they work

The larynx, particularly the action of the vocal cords, is the crucial element in voicing. Let us take a look at an example. Think of when you want to warm your hands in cold air using your breath: you basically spread your vocal cords to leave the glottis open for air to flow freely into and eventually out of your mouth. This is the way the sound [h] in the word hot is pronounced. And this is precisely why the English sound [h] is called voiceless because in its production, the vocal folds are wide open and do not vibrate.

However, if the vocal cords are brought close together narrowing the glottis, the passage of air will set the vocal cords into vibration creating an audible “noise”. Press your fingers on your larynx (the bony structure in your throat) as you pronounce a stretched out [zzzzzzzzzz]. Do you feel the “buzz”? It is the vibration of the vocal cords that creates this buzzing sound. The English sound [z], as in the word zoo, is thus labeled voiced. Some other voiced sounds in English are [b] as in bee, [g] as in go, [ð] as in thee, [ʒ] as in measure, and all of the vowels (we will discuss these sounds in more detail below). Contrast these with sounds like [s] as in see and [ʃ] in shhh! which, like [h], are classified as voiceless as they are produced with an open glottis and therefore no “buzz”.

Box (1): Letters and Sounds As you have noticed above, phoneticians (linguists specializing in the study of speech sounds) use special symbols often placed inside brackets to refer to a certain sound (e.g. [dʒ] for the initial and final sounds in “judge”). Letters of the alphabet can be very confusing, even misleading, in their representation of sounds. Take the case of “c” in critic, cello and ocean” or “ch” as in “chic, chemistry and China”. It is easy to see how the same letter(s) could stand for different sounds. That is why phoneticians have, over the years, come up with a system, called the International Phonetic Alphabet (IPA for short), in which each symbol stands for one and only one sound. Therefore, for example, “ch” in chic is shown as [ʃ], “ch” in chemistry is written as [k], and “ch” in China is indicated by [tʃ] ***Can you assign the right symbols for “c” in “critic, cello and ocean”?*** Note also that placing these symbols inside brackets, as phoneticians do, helps readers recognize that one is talking about sounds and not letters of the alphabet. It is therefore essential to “keep the notions of ‘letter’ and ‘sound’ distinct in your thinking about pronunciation” (Gussenhoven & Jacobs 2017, p. 1).

Click here for the pronunciation of each speech sound on the IPA chart 

As we just observed, sounds can be produced as far back as in the larynx. Let us look at another example of a glottal sound (one produced in the glottis). Try and say uh oh!, the pair of sounds produced when one makes a mistake. For this sound to be pronounceable, the vocal cords need to be closed off completely, allowing pressure to build up from the air coming from the lungs, and then suddenly opened to release the air. This sound, shown with the IPA symbol [ʔ] and sometimes found in the pronunciation of “t” final as in cut, is referred to as a glottal stop or plosive - “stop” indicating the stoppage of air and “plosive” its sudden release, or plosion, during the articulation of [ʔ]. Coughing is another way of producing a forced glottal plosive although it is not considered a speech sound!

This image captures the various configurations of the vocal folds during speech

Box (2): Place vs. Manner of Articulation As you have just noticed, when describing the production of consonant sounds, phoneticians mention the “place” where the sound is articulated (e.g. a glottal sound is one produced in the glottis) and the “way” it is articulated (e.g. whether there is any obstruction of the airflow). Thus, a consonant sound can be described and classified in terms of its “place of articulation” and also its “manner of articulation”. A -glottal stop- is, therefore, a sound that is produced with an obstruction and sudden release of air in the glottis. Note these place and manner distinctions as you read through the rest of this section.

To summarize:

• The action of the diaphragm and the lungs make respiration possible, a crucial aspect in the production of speech.

• During exhalation, the air travels up the trachea and passes through the larynx.

• “Voicing” is a general feature relating to the vibration - or lack of it - of the vocal cords.

• A voiceless glottal stop [ʔ] can be produced by shutting off the vocal cords and suddenly opening them to release the built-up air.

----------------------------------------------------------------------------------------------------------------------

Consonants

So far, we have looked at the two basic mechanisms of respiration and phonation in speech. We also discussed the production of one sound, the glottal stop [ʔ], in the larynx. But articulation is a process that mainly occurs above the larynx, technically referred to as the vocal tract. The vocal tract in humans consists of two cavities: the oral cavity, the mouth, and the nasal cavity, the nose. It is the filtering of the airstream in the oral and nasal cavities and the action of the articulators (organs such as the tongue and the lips) that determines the final shape of a speech sound.

See a picture of the vocal tract here

In this section, we will be following the flow of air out of the larynx and into the vocal tract. Along the way, we will discover the various sounds (mostly of English) and their places and manners of articulation. Following Catford 2001, I would like to encourage you to use any tips or tricks mentioned to become aware of and to observe introspectively where and how various sounds are produced as well as any similarities and differences in their place and/or manner of articulation.

Air flowing out of the larynx first enters the pharynx, a muscular tube in the upper part of the throat. What is often called a “gag reflex” happens in this region of the throat. Pharyngeal sounds, i.e. sounds made in the pharynx, are generally alien to English but are common in Arabic and middle-eastern Hebrew. Take the example of a throaty [h] sound represented by ح in Arabic (the IPA symbol for this is [ħ]): the pronunciation of [ħ] is achieved by moving the root of the tongue close enough to the back wall of the mouth to narrow the passage of air, thus creating a pharyngeal consonant sound. Since there is no vibration of the vocal cords accompanying the production of this sound, [ħ] is classified among voiceless consonants.

Moving further up from the pharynx, one can easily see a fleshy extension hanging above the throat. This structure is called the uvula (see image link above to locate this). Uvular consonants are sounds articulated with the back of the tongue close to or against this structure. Perhaps the most well-known example of this would be the voiceless uvular “r” (shown mirror image of capital R [ʁ] in IPA) found in words like Louvre and Paris in standard French.

Box (3): Source-Filter Theory Before moving further, it is useful to point out that the oral cavity, as you have seen so far, is made up of flexible “active articulators”, such as the tongue, that interact with other articulators, some of which relatively immobile “passive articulators”, to produce the sounds of speech. The action and positioning of these various articulators modifies the shape of the oral cavity, creating a resonating chamber for the formation of various sound qualities. A well-accepted theory in phonetics, known as the “source-filter theory”, posits that it is this action of the articulators in the vocal tract on a phonation source that is responsible for the production of sound (see Seikel et al. 2009, p. 269 for a complete discussion).

Past the uvula and moving forward on the roof of the mouth, one reaches the velum, or the soft palate. It is easy to touch this surface by arching the back of your tongue against the back roof of the mouth. This is where the velar consonants [k], [g] and [ŋ] (as in king, guy, and sing respectively) take shape. Try and assume the tongue position for the three sounds above. Do you notice the tongue sticking to the velum to stop the flow of air? Since there is a blockage and sudden release of air in the manner [k], [g] and [ŋ] are pronounced, each of these is called a velar stop.

Question (1): Which of the above three sounds is voiceless? Which is a nasal sound (i.e. produced by allowing the air to escape through the nose)? [See the table (1) further below to check your answers]

Now, try pronouncing the sounds [ʃ] as in shhh! (when “shushing” someone) and [ʒ] as in measure. You will notice that these two sounds share roughly the same place of articulation, being articulated further ahead in the mouth than [k] or [g], on the hard palate (the center of the mouth roof). Due to this, these sounds are often labeled palatal.

Now, another difference you’ll notice is, unlike [k] or [g], for [ʃ] and [ʒ] you do not block the passage of air in your mouth but rather restrict it, creating so-called “friction” between the tongue and the hard palate, which is why [ ʃ ] and [ʒ] are called fricatives in terms of their manner of articulation.

There are other consonants that can be produced at the palatal (also known as post-alveolar) point, although their manner of articulation differs from the above. One class represents the initial and final sounds [tʃ] in church and [dʒ] in judge in English.

The case of [tʃ ] and [dʒ] is interesting: in fact, they both start as a stop, [t] and [d] (see discussion of these sounds further below), but gradually turn into a fricative, [ʃ] and [ʒ] respectively. That is why [tʃ] and [dʒ] are placed in a different class for manner of articulation: affricates, consonants that are a cross between a stop and a fricative.

Question (2): Which of the affricates above would you say is voiceless? Which is voiced? [See the table (1) further below to check your answers]

Another interesting case of a post-alveolar consonant is English [r] as in road. Although the tip of the tongue raises to touch the roof of the mouth, air can still flow from the sides of the tongue. Also, note that the tip of the tongue is slightly curled up toward the hard palate (the reason why [r] is sometimes called a retroflex) and the sides of the tongue come in touch with the back teeth. Due to the fact that air can pass through the sides of the tongue in the production of this speech sound, [r] has been labeled lateral (lateral meaning “from the sides”).

Moving past the hard palate, most people can easily locate with the tip of their tongue a protruding bony structure behind the teeth called the alveolar ridge. Pressing the tip of the tongue against the alveolar ridge will block the flow of air that is compressed behind the closure. Releasing this airflow will result in two fairly common consonants in English: [t] as in tick and [d] as in day. That is why [t] and [d] are labeled alveolar stops. Note that like many of their paired counterparts above ([k] & [g], [ʃ] & [ʒ], [tʃ ] & [dʒ]), one in this pair, i.e. [t], is voiceless while the other, [d], is voiced.

However, [t] and [d] are not the only alveolar stops. There is one more alveolar stop in English: the consonant [n] as in nine. The reason we are treating this separately is because in order to pronounce [n], the airflow gets redirected and released through the nasal rather than the oral cavity, a feature it shares with another stop, [ŋ], that we covered further above. This is why [n] is considered a nasal (alveolar) stop.

Box (4): Oral vs Nasal Let us take the word “noon”. Now, try saying the word while you pinch your nose. It is impossible to say noon while your nose is blocked (similar to when you have a cold). In fact, the word becomes more like “dude” in pronunciation! It is easy to realize the difference: while [d] and [n] share the same place and manner of articulation, they are different in that [d] is pronounced with the air passing through the mouth - or the oral cavity (therefore called an “oral” consonant) and [n] with the air passing through the nose - or the nasal cavity (thus labeled as a “nasal” consonant).

By the way, stops are not the only consonants articulated at the alveolar point. Just in the case of [ʃ] and [ʒ] that we covered earlier, it is possible to bring the tongue close to the roof of the mouth but not touching. This would allow the air to escape with a “friction,” resulting in consonants that we called fricatives. There are two alveolar fricatives in English: [s] as in space and [z] as in zoo. 

Question (3): Which of the two alveolar fricatives is voiced? Which is voiceless?

Fun Fact (1): Unlike stops and affricates, fricatives can be continued until one runs out of breath!

Fun Fact (2): Consonants that share the same place of articulation are called homorganic ("homo" meaning same, and "organic" referring to the speech organs).

Wait! We are not done with the alveolar ridge yet, which actually shows how important it is as a place of articulation in English. According to Dewey (1923), “alveolar sounds account for almost 50% of the sounds in English” (Kent 2017, p.30). Three more sounds, albeit with different manners of articulation, are produced at the alveolar ridge. One is the consonant [l] as in little. However, note that although the tongue tip sticks to the alveolar ridge in the pronunciation of [l], air is allowed to escape from both sides of the tongue. One way to test this is to hold the tongue in place for the pronunciation of [l] and suck air in: this way, you can feel the cold air passing on both sides of your tongue. For this reason, [l] is considered another example of a voiced lateral.

Another particular alveolar consonant is a so-called tap which is characteristic of a mostly North American pronunciation of the middle [t] sound in words like butter and matter (shown with the IPA symbol [ɾ]). To articulate a tap, the tongue tip is raised to make a quick “hit” against the alveolar ridge while moving back against the roof of the mouth.

A similar consonant, called a flap (as in a typical North American pronunciation of [t] in party and often shown with the IPA symbol [ ɽ ]), uses the same procedure, but this time the tongue hits “forward” (with the tongue initially positioned behind the alveolar ridge) instead of “backward” (as in the case of a tap).

As air from the lungs moves forward in the oral and nasal cavity, the articulators find novel ways of obstructing its flow to create a variety of consonant sounds. In English, the last point of involvement of the tongue as a primary active articulator is its contact with the upper front teeth to restrict the flow of air. This is the manner in which the dental fricatives [θ] as in three and [ð] as in thee are produced. Note the respective voiceless vs. voiced distinction between [θ] and [ð].

The next two places and manners of articulation do not involve the tongue in a major way but enlist the lips and/or teeth as active articulators. The first group of these involve the lower lip and the upper front teeth in constraining the airflow to create labio-dental fricatives. These include the consonants [v] as in vein and [f] as in feign in English, with [v] being the voiced and [f] being the voiceless counterparts in this class.

The final point of closure is between the lips. By pressing the upper and lower lips together, it is possible to momentarily seal off the airstream from the lungs before opening them to release the built-up pressure in the oral cavity. This is the mechanism by which the bilabial plosives [p] and [b] (as in pan and ban respectively) are articulated.

Question (4): Can you tell which in this pair is voiceless and which is voiced?

Another bilabial consonant is [m] as in man. It is easy to see that [m] is a nasal stop as, in order to realize its pronunciation, the air needs to escape through the nasal cavity. Try enunciating the word mom while pinching your nose: it becomes a totally different word: Bob! (Denham & Lobeck 2013, p.81). Note that [m] is a voiced consonant.

Box (5): Consonants vs. Vowels (and the case of [w] and [j]) We have used the term consonant quite a few times in describing the speech sounds above. But what is a consonant? One basic distinction often drawn in the description of speech sounds is between “consonant” and “vowel” sounds. As we have observed so far, consonants are speech sounds that are produced by the obstruction of the air flow through the vocal tract. For instance, [p] and [g] can only be pronounced through a complete stoppage of the airstream while [s] and [v] are produced when the obstruction is partial, allowing the air to escape through a narrow passage. Vowels, which we will discuss further below, are produced when the flow of air is not obstructed. Instead, changes in the shape of the oral cavity and the configuration of the tongue and lips determine the final product (e.g. the sound [o] in go). Yet, there are consonants whose obstruction of the airstream is so small that they end up exhibiting some of the features of vowels (e.g. the initial bilabial [w] and palatal [j] sounds in with and yet, respectively). These consonants are termed “approximants” and they are always followed by vowels.

Table (1) below summarizes all of the consonants we discussed above in terms of their place and manner of articulation as well as voicing (click table to enlarge):

One interesting thing coming out of this table is the ease with which we can describe consonants like linguists do. Combining voice + place + manner of articulation will help describe a consonant: e.g. [d] would be described as a voiced alveolar plosive while [t] would be a voiceless alveolar plosive. It also helps us see how similar or different two or more consonants are.

Try describing 1) [k] vs. [g], 2) [s] vs. [z], and 3) [v] vs. [f] using the voice + place + manner arrangement. What similarities and differences do you notice between the sounds?

To summarize:

• Once air is pushed out of the lungs and through the larynx, the action of the various articulators in the vocal tract modulates this airflow to create the majority of sounds in the world’s languages.

• Consonants are speech sounds produced at certain points of obstruction in the vocal tract. This obstruction can be complete, as in the case of plosives and affricates, or partial, as in the case of fricatives.

• Consonants are therefore described in terms of their: a) place of articulation, b) manner of articulation, c) voicing.

• Nasal consonants are speech sounds, such as [m], [n] and [ŋ], that are produced as a result of air flowing through the nasal rather than the oral cavity.

----------------------------------------------------------------------------------------------------------------------

Vowels

Vowels, unlike consonants, are not discussed in terms of place or manner of articulation as there is virtually no contact between pairs of articulators - and therefore no obstruction of the airflow - in the production of vowels. However, phoneticians use three major criteria to determine vowel quality:

1. Frontness (whether they are pronounced toward the front, mid or back of the oral cavity), 
2. Height (the relative tongue position in the oral cavity: high, mid or low), 
3. Rounding (whether the lips are spread or rounded while producing a vowel sound).

As you can see, voicing does not figure in the above criteria. This is because all vowels are voiced, so it would be redundant to mention it in vowel description. Let us consider a few examples of vowels below and then look at a diagram of the vocal tract in the articulation of vowels in English.

Compare [i] (as in read), [ɑ] (as in father), [u] (the vowel in rude): The first thing you will notice while trying to pronounce these sounds is the absence of obstruction of the air flow in the oral cavity. But where in the mouth these sounds are pronounced certainly varies, with [i:] being more in front with the lips spread (or unrounded), and [u:] further in the back with the lips rounded. You will also notice that for [i:] and [u:], the tongue is in a rather high position in the mouth compared to [a:] for which the tongue position is low (physicians often ask patients to say [a:] as the lowered tongue will allow better visibility of the back of the mouth).

See this image for the vowel sounds above and a few others

You can also notice the movement of the jaw, another active articulator, along with degrees of lip spread or rounding by repeating the sequence he-who-ha. English has a dozen of these sounds, some of which are a combination of two or even three vowels: consider, for example, [aʊ] in shout, a sequence of two vowels, and even a sequence of three vowels [aʊə] as in hour (the former is often called a diphthong and the latter a triphthong).

English vowel quadrilateral depicting vowel frontness, height & rounding with examples

Just as place and manner descriptors are used to identify consonants, phoneticians use the above features to describe vowels. For example, [i] would be a high front unrounded vowel while [u] would be described as a high back rounded vowel. Note also that vowels are important because, unlike consonants, they form the nucleus of syllables (a topic we will explore in another post).

To summarize:

• Vowels, unlike consonants, do not obstruct the flow of air in the oral cavity. That is why place and manner of articulation would be irrelevant to the description of vowels.

• Vowels are, therefore, described in terms of three criteria: a) frontness, b) height, c) rounding.

• Vowels are always voiced (except for certain phonetic environments that cause devoicing: a topic of discussion in phonology).

----------------------------------------------------------------------------------------------------------------------

At this point, you might wonder, why is there so much emphasis and detailed analysis of every single speech sound? This is because phonemes, the word phonologists use to refer to distinct speech sounds, are the minimal elements that represent and distinguish meaning (Kent 2017, p.8). Let me clarify with an example. Compare the words map and nap: they are identical in their vowel and final consonant sounds [-æp]. What distinguishes these as two different words with different meanings in the English language is the difference in their initial phoneme, [m] versus [n]. And further note that these two phonemes are different only in their place of articulation (in manner and voicing, they are the same!). As such, small phonemic distinctions create differences in words and their meanings in the world's languages. Identifying and accurately describing them is, therefore, a first step in linguistics, the science of language.

However, note that speech sounds are rarely produced in isolation. Everyday speech comprises utterances that are made up of strings of sounds arranged in a linear fashion. Neighboring sounds affect the qualities of one another in multiple ways. There also are phonetic features that extend beyond a single segment of sound and affect syllables and sentences, a topic we will return to in a next post on articulatory phonetics.

-----------------------------------------------------------------------------------------------------------------------

References

Catford, J. C. (2001). A Practical Introduction to Phonetics (2nd ed.). Oxford: OUP.

Collins, B. & I. M. Mees (2013). Practical Phonetics & Phonology: A Resource Book for Students (3rd ed.). London: Routledge.

Denham, K., A. Lobeck (2013). Linguistics for Everyone: An Introduction (2nd ed.). Boston: Wadsworth, Cengage Learning.

Dewey, G. (1923). Relative Frequency of English Speech Sounds. Cambridge: Harvard University Press.

Gussenhoven, C. & H. Jacobs (2017). Understanding Phonology (4th ed.). London: Routledge.

Kent, R. (2017). Normal Aspects of Articulation, in J. E. Bernthal, N. W. Bankson, P. Flipsen Jr. (eds.), Articulation & Phonological Disorders: Speech Sound Disorders in Children (8th ed.), pp. 7-48. NJ: Pearson Education.
 
Seikel, J. A., D. W. King, D. G. Drumright (2010). Anatomy & Physiology for Speech, Language & Hearing (4th ed.). NY: Cengage Learning.