0. Introduction

Dragons have a vocal tract that is some respects similar to that of humans, but with notable differences that affect language capabilities in many ways, including their inventory of possible phones (speech sounds). In this article I collect modern scholarship into an overview of the state of the field, with particular reference to the data from my own fieldwork documenting Xa:lmissan. As the data collection process for draconic phonetics is arduous, the discoveries piecemeal, and the researchers occasionally pieces of meals, despite its brevity this article represents a hard-won advance for the discipline.

In section 1 I review elements of the draconic vocal tract that affect possible speech sounds and general draconic phonetics. The focus is on points where the dragon’s anatomy differs from human anatomy enough to have a strong influence on speech (e.g. the internal structure of the nasal tract and the larynx are not discussed, because they largely correspond to the human equivalents). See Figure 1 for a diagram of the draconic vocal tract. In section 2 I set out the symbols of the Draconic Phonetic Alphabet. Section 3 concludes.

1. Draconic Physiology and Vocal Tract

1.1 Oral Tract Length and Shape, Lips, and Teeth

In every known species, consonants differ from vowels in that consonants involve some degree of constriction, or blockage of the airflow out from the lungs. The snout of the dragon is much longer and narrower than the human mouth, which, in combination with the more flexible tongue, allows for several additional points of contact for such blockage where consonants unavailable to humans may occur (discussed in 1.2 below).

For vowels, however, the long narrow mouth is constraining. While the human mouth allows for low, mid, and high vowels, dragons’ mouths admit only two degrees of height, low and high; there simply isn’t enough vertical space for distinguishable mid vowels.

In humans, vowels may be affected by lip rounding, i.e. whether the lips are pursed together or held flat (consider the lip shape when saying “flow” vs “flee”). Dragons lack lips altogether, meaning there is no possibility of rounded vowels. Possibly due to this lack of rounding, even though the oral tract is long, high vowels are generally divided into only four degrees of backness: front, mid-front, mid-back, and back. Anything further would likely be indistinguishable. Low vowels are either front or back. In total, then, most draconic languages have a maximum of only six vowels.

Art by Leah Nodar, ink/colors by Tim Shirley. Based on Figure 2, “Proposed vocal tract configuration of a Chinese alligator during bellowing,” in Reber et al (2015).

Perhaps to compensate for the small number of possible vowels, draconic languages often use vowel voice quality phonemically. Changes in voice quality involve speaking in ways that modify the basic voice with an additional texture, such as raspiness. The three main voice qualities so far documented among dragons are breathy voice, creaky voice, and smoked voice (discussed in 1.3 and 2.2 below).

Because dragons lack lips, they have no consonants that involve lip mobility, which occur frequently in human languages. Arguably /m/ is the only labial consonant clearly present in draconic languages (though see also 2.1 below). Also, because their teeth are obscenely sharp, consonants made by touching the lips or tongue to the teeth (as in the first sounds of “fee” and “thee” respectively) are not viable.

1.2 Tongue: Pharyngeal Ridge and Frenulum

Two key differences in the tongue affect draconic speech. The first is the presence of the pharyngeal ridge at the back of the tongue, hypothesized to protect the tongue during fire-breathing. This knobbly bone ridge has no counterpart in the human mouth. In humans, the root of the tongue can be wiggled backward towards the far side of the throat, as is found in certain consonants in Arabic and other languages (Zsiga 2013). Although dragons have an enormous pharynx in comparison with humans, the pharyngeal ridge means that there are no such pharyngeal consonants (Putterill and Soley 2003, Putterill and Soley 2004, Reber et al. 2015).

The second difference is the much shorter length of the frenulum in comparison to the overall length of the tongue. The frenulum is a small piece of tissue that attaches the tongue to the base of the mouth. In humans the frenulum makes this connection about a quarter of the way down the underside of the tongue (Zsiga 2013:87); in dragons the frenulum is more than halfway down the underside of the tongue. This means that the draconic tongue blade is far longer and has greater mobility. In combination with the greater overall length of the oral tract, this allows for three places of articulation that are not possible in humans: retro-velar, dorso-velar, and dorso-retroflex. Diagrams of these tongue positions are given in Figure 2.

Art by Leah Nodar, ink/colors by Tim Shirley.

In humans, retroflex consonants are made by curling the tongue backwards so that the underside touches just behind the alveolar ridge, which is the line of gums that holds the front teeth. Such consonants are common in Indian languages, such as Tamil and Hindi. Draconic retro-velar consonants continue that backwards curling until the underside of the tongue tip reaches all the way back to the soft palate, near the back entrance to the nose. The consonants usually called retroflex are in this paper called retro-alveolar, to maintain the distinction with retro-velar.

For the dorso-alveolar and dorso-retroflex consonants, the dragon’s tongue touches the roof of the mouth in two places simultaneously. In both cases the base or body of the tongue moves up towards the soft palate; the blade of the tongue dips down towards the lower jaw and then back up again, with the tip pointing forward for dorso-alveolar or curled back for dorso-retroflex. Human tongues are simply neither long enough nor flexible enough to come anywhere near these articulations.

Dorso-alveolar and dorso-retroflex placements are generally fricatives, approximants, or nasals. Dorso-alveolar or dorso-retroflex obstruents are unattested, which is unsurprising as they would be indistinguishable from velar obstruents. Theoretically, it seems like the pocket of air created in the space between the two tongue-roof constrictions could be rapidly expelled for a click. However, neither this nor any other click has so far been attested in a draconic language. (Gernie Handervoss won the prestigious Pellenfarth Grant in 2014 for an ingenious experiment designed to determine whether such clicks were unpronounceable or simply unattested, but she was unfortunately eaten at the field site.)

1.3 Combustal Pipe and Combustal Flap

Naturally, the most distinctive characteristic of dragons is the ability to breathe fire. While the full mechanism of fire-breathing is beyond the scope of this paper, a few points relevant to draconic phonetics are considered (drawn from Sooner 2022). The combustal pipe runs below the windpipe, and is a narrow, tough tube, ending at the pharynx in the spark ring and the combustal flap. The combustal flap covers and protects the spark ring and combustal pipe when eating. The spark ring is a ring of muscle surrounding eighteen thin slats of semi-metallic bone. When the spark ring muscles are contracted in quick pulses, the slats hit against each other, producing sparks. In fire-breathing, the combustal flap is pulled down, and a fine mist of suspended particles similar to methane is propelled down the combustal pipe towards the pharynx, lit by sparks from the contraction of the spark ring, and fed by oxygenated air moving up the windpipe from the posterior air sacs (see Figure 3 and discussion of pulmonic system in section 1.4). The important point for draconic phonetics is that the combustal flap may also be pulled back during speech, in which case air from the air sacs flows over the spark ring as it is expelled, producing a hollow noise similar to the sound made by humans blowing over the top of an empty bottle. Speech sounds produced with the combustal flap pulled down are referred to as having the “smoked” voice quality, most often smoked vowels (Falsworth et al. 2018).

Of secondary importance, but worth noting, is that the epiglottis equivalent in dragons is two flaps attached at the sides of the throat, which work together to cover the larynx during swallowing. They also pull the windpipe down, expanding the opening to the esophagus and allowing the dragon to eat large creatures in a single gulp, a task abetted by the extremely large pharynx (compare this video of a crocodile swallowing a gazelle).

1.4 Unidirectional Breathing

Dragons have a unidirectional pulmonic system, analogous to that of birds and crocodiles (Schachner et al. 2013, Song n.d., Wedel 2013). In humans, the lungs do two jobs: they move air through the body, and they allow for the exchange of oxygen and carbon dioxide. In a unidirectional system, these two jobs are split up. When breathing in, the posterior air sacs expand, creating low-pressure areas that air moves into. These air sacs then contract, forcing air through the parabronchial lungs (a one-way system) and into the anterior air sacs. For exhaling, the anterior valve shuts, preventing air full of carbon dioxide from reentering the posterior sacs.

The posterior sacs are split into two sections, the thyroidal and the abdominal, which allow free passage between them during normal respiration. During fire-breathing, the intra-sac valve between the two closes them off from each other, the abdominal valve closes, and the thyroidal valve opens, so that the thyroidal air sac can move oxygen-rich air down the windpipe to feed the fire, while the abdominal air sac continues to move oxygenated air into the parabronchial lungs. See Figure 3 (Song n.d.).

Art by Leah Nodar, ink/colors by Tim Shirley. Based on "Avian Respiration" in Song (n.d.).

This system also has an important effect on draconic speech at the discourse level. The air sacs take up a great deal of space within the dragon’s torso and hold a far higher volume of air than human lungs. Thus, while a human can say about one long clause before running out of breath, the usual length of an utterance in a dragon’s speech is closer to what a human would consider a paragraph.

As a morphological side note, Xa:lmissan and other draconic languages use arbitrarily many discourse-referent pronouns. That is, it is common to find systems where pronouns refer back to when each individual was first mentioned in the current discourse. To use an artificial example: “Jim came to meet Bob. I watched ‘the first one’ (=Jim) talk to ‘the second one’ (=Bob).” These pronouns are modifications of the number system, and so can go on indefinitely; because the average length of a dragon’s utterance is so long, the number of such referents regularly climb into the dozens. Dragons do not seem to have any difficulty keeping track, and easily parse a sentence such as “I watched ‘the seventeenth one’ and ‘the twenty-third one’ talk to ‘the fourth one.’” Human interlocutors, however, tend to be quickly overwhelmed.

2. DPA (Draconic Phonetic Alphabet)

2.1 Consonants

The above table consists of the consonants that researchers have so far attested in any draconic language. Some of these remain a matter of considerable debate. Evidence of /m/ as a phoneme has been found in enough languages (including, obviously, Xa:lmissan) that it is no longer questioned, but /p/ and /b/ have been attested to only as allophones, and only rarely. Pumblecrum (2012) argues that the common draconic phonological pattern of word-final lenition to nasals is in itself somehow evidence that word-final /m/ in Xa:lmissan is underlyingly a historical /b/ or /p/. This researcher, however, agrees with Zelen et al (2002) that such postulations are wishful thinking.

As with clicks, taps and flaps have yet to be attested in a draconic language, though some  believe this is more likely a failure of instruments than a true lack. After all, no one has yet managed to persuade a dragon to make high-quality recordings for spectrographic analysis.

2.2 Vowels and Voice Quality

As discussed above, the long, narrow vocal tract of the dragon, though it adds a variety of consonants unavailable to the human mouth, greatly shrinks the available vowel space. Instead, voice quality often plays an important role in increasing the number of distinguishable vowels.

Breathy voice involves keeping the vocal folds open while they are vibrating, which adds additional air from the lungs for a whispery quality; creaky voice involves slowing the vocal fold vibrations for a rasping quality (Thomas 2016); and smoked voice, described in 1.3 above, adds a deep, hollow quality. Since both breathy and creaky voice originate at the glottis, but smoked voice originates at the combustal flap, it is possible for a vowel to be pronounced simultaneously smoked and creaky, or smoked and breathy (but not breathy and creaky, nor all three at once). Some languages allow the combinations smoked-creaky and/or smoked-breathy as their own separate voice qualities, for up to a six-way distinction.

Pumblecrum (2015) argues for theoretical reasons this must be mistaken, and that a full six-way system of voice quality distinction is unsupportable; and should Pumblecrum ever summon up enough courage to leave his office and travel to Xa:lmɔƕƹɔ himself, I believe that at that time his opinions as to the accuracy of other people’s transcriptions will deserve attention.

As when transcribing human phonation, breathy-voiced sounds are marked with a diaerisis below, e.g. [ɔ̤] and creaky-voiced sounds with a tilde below, e.g. [ɔ̰]. The dragon-exclusive addition, smoked-voiced sounds, are marked with a double-tilde above, e.g. [ɔ͌].

3. Final Comments

The field of draconic phonetics and phonology has made great strides in the last two decades, and the basic groundwork has at long last been laid, though at great cost. As per tradition, I end with a salute to those brave scholars who have given their lives for the field, and add to that list this month’s departed: J. Julian (University of Ekoarak); Christine Dockrill (Phillips College); Nytan Taylor (independent researcher); and Raymond Baib (Center for Applied Draconic Sciences). May they rest in peace.

Leah Nodar

4. Works Cited

Falsworth, J., Roberts, M.N., Alex, P., and Steffs, L. C. Voice quality in draconic languages. Mythological Phonetics, 18, 21-44.

Pumblecrum, Q. F. (2012). Labial obstruents revisited: Evidence from historical Xa:lmissan. Journal of Draconic Languages, 42, 95-110.

Pumblecrum, Q. F. (2015). Resolving complex consonant clusters: Assimilation or deletion? In C. P. Pieken and A. Quesada (Eds.), Sound patterns of dragons: New perspectives on phonology (25-45). University of Ekoarak Press.

Putterill, J. F., and Soley, J. T. (2003). General morphology of the oral cavity of the Nile crocodile, Crocodylus niloticus (Laurenti, 1768). I. Palate and gingivae. Ondespoort Journal of Veterinary Research, 70, 281-297.

Putterill, J. F., and Soley, J. T. (2004). General morphology of the oral cavity of the Nile crocodile, Crocodylus niloticus (Laurenti, 1768). II. The tongue. Ondespoort Journal of Veterinary Research, 71, 263-277.

Reber, S. A., Nishimura, T., Janisch, J, Robertson, M, and Fitch, W. T. (2015). A Chinese alligator in heliox: formant frequencies in a crocodilian. Journal of Experimental Biology, 218, 2442-2447.

Schachner, E. R., Hutchinson, J. R., and Farmer, C. G. (2013). Pulmonary anatomy in the Nile crocodile and the evolution of unidirectional airflow in Archosauria. PeerJ, 1, e60.

Song, R. No date. Avian Respiration. http://people.eku.edu/ritchisong/birdrespiration.html

Sooner, M. P. (2022). Anatomy of the fire-breathing apparatus in north Kalaha dragons. Dragonfire, 75, 19-27.

superj72. (2012). CROC SWALLOWS GAZELLE WHOLE!!! YouTube. https://www.youtube.com/watch?v=pWo4z2v7KhY&t=100s

Thomas, Erik. (2010) Sociophonetics: An introduction. Palgrave Macmillan.

Wedel, M. (2013). Unidirectional airflow in the lungs of birds, crocs… and now monitor lizards!? Sauropod Vertebra Picture of the Week. https://svpow.com/2013/12/11/unidirectional-airflow-in-the-lungs-of-birds-crocs-and-now-monitor-lizards/

Zelen, S., Starshir, B., Sheu, V., and Starshir, S. A. L. (2002). Empirical issues in auditory processing of labial consonants. Journal of Draconic Languages, 22, 39-64.

Zhao-Xian, W., Sun, N. Z., Mao, W. P., Chen, J. P., and Huang G. Q. (1991) The breathing pattern and heart rates of Alligator sinensis. Comparative Biochemistry and Physiology, Part A: Physiology, 98(1), 77-87.

Zsiga, E. C. (2013). The sounds of language: An introduction to phonetics and phonology. Wiley-Blackwell.