When people find out I’m a phonetician who happens to be interested – among other things – also in the potential effects of the menstrual cycle on language and speech, they typically present me with fairly excited reactions. Why would a linguist be interested in anything to do with the menstrual cycle? I’ll attempt to answer this question in this blog entry, with the caveat that my research focuses on phonetics and phonology, i.e. sounds and not, for instance, word order.
Larynx as a sexual organ
The larynx has been called a secondary sexual organ (e.g. Abitbol et al. 1989; Amir & Biron-Shental 2004; Collins & Missing 2003; Hall 1995; Henton & Bladon 1985). This is because the larynx does a considerable amount of work when it comes to signalling interpersonal relationships by producing a wide range of voice qualities* with a wide range of functions. We can take breathiness as one example to explore. Breathy voice is a voice quality associated with more air passing through the vocal folds when speech is produced – have a listen to Marilyn Monroe’s “Happy birthday, Mr President”, or a more recent version by Lana Del Rey, who’s also fairly breathy here. Breathiness in female speech has been traditionally associated with attraction: the breathier a woman’s voice sounds, the more attractive the voice gets rated by heterosexual male listeners (e.g. Liu & Xu 2011). In addition, female voices seem more attractive to male listeners when the female they listen to happens to be most likely to conceive (Pipitone & Gallup 2008: 271). These perceptual results are phonetically grounded: production studies show that female voices can indeed change depending on the phase of the menstrual cycle (Raj et al. 2008). Why should this be the case?
The larynx is a fairly complex organ, which composes of an array of muscles, cartilages, and other tissues (see e.g. Moisik 2013). But a picture is worth a thousand words, so check out the complexities of my own larynx in the picture to the left. As you can see, the larynx is indeed complex and it is also a somewhat wet place. See also the picture to the right, which shows a bit more saliva.
The physiology of these different structures can be affected by a range of factors, including hormonal changes. Estro-progesterone changes associated with different phases of the menstrual cycle may lead to the thickening of laryngeal mucus, reduction of hydration of the vocal fold edges, and oedemas connected with reduced lubrication (e.g. Abitbol et al. 1999; Pipitone & Gallup 2008; Raj et al. 2008), and thus affect the voice quality produced.
Hormonal effects on overall voice quality, or also contrastive sounds?
Considering that menstrual hormonal changes can impact the physiology of the laryngeal structures sufficiently to lead to voice quality differences, I asked myself the following: what about laryngeal phenomena that contribute to a contrast implementation, such as aspiration or voicing of obstruents (i.e. fricatives – such as /f/ or /s/, as in fish and sea; plosives – such as /p/ and /k/, as in punk and pumpkin; and affricates – such as /tʃ/ and /dʒ/, as in chocolate and jazz)? If voice quality can get affected by hormonally induced changes in the laryngeal structures, why shouldn’t smaller properties of language whose articulation relies primarily on these laryngeal structures, such as aspiration and voicing, be also affected?
Interestingly, Whiteside et al. (2004) and Wadnerkar et al. (2006) were interested in similar questions. They measured Voice Onset Time in /p, t, k/ and /b, d, g/, using English data. To simplify, Voice Onset Time (VOT) enables us to see to what extent a consonant such as /p/ or /b/ is aspirated or voiced. What Whiteside et al. (2004) and Wadnerkar et al. (2006) found was that the VOT difference between the /p, t, k/ and the /b, d, g/ groups was magnified “at the high hormonal phase of the menstrual cycle” (2006: 21) by 1.39-3.51 milliseconds.
I decided to conduct a similar study, focusing on consonants that would allow me to deal with laryngeal phenomena of different phonological relationships. So, I wondered whether, for instance, what’s known as allophonic variation in voicing may be more susceptible to potential menstrual hormonal effects than phonemic variation. Let me explain what I mean by allophonic variation with an example from Czech. The Czech consonant /ř/ is supposed to have two variants, or allophones: a voiced one and a voiceless one. We can predict where we get which: if we find the sound in the vicinity of a voiceless sound, then it is voiceless itself (as in tři “three”, where we get the voiceless variant; as opposed to dři “toil/drudge!”, where we get the voiced one). Because these two /ř/ variants are allophonic though, replacing the voiced and the voiceless versions doesn’t change the meaning of the word, as might happen with the English Tomand Domin case of VOT. In addition, I also wondered about the methods one should use to investigate the questions I became interested in. Out of curiosity then, I recorded myself producing the same thing every day for a period of eight months, keeping myself – to my own surprise – wonderfully ignorant as to how many phases the menstrual cycle has and how long these last**. The results of this experiment were never intended to be published. However, after some of my colleagues insisted on learning about if I was finding any consonantal variation predictable by menstrual phase, I succumbed to this peer-pressure in hope this interesting topic would move closer to the spotlight of linguists and did indeed decide to publish the results (Hejná 2019/Forthcoming).
And what did I find?
First of all, I found that the methodological beasts associated with the task are not easily slain without opening a can of worms regarding how to establish, for instance, when ovulation takes place if you also want to do your best to keep your subject as ignorant on the topic as possible. Using less reliable post-hoc methods to establish when I was ovulating, some rather interesting results nevertheless emerged:
- My voice quality was the breathiest during ovulation, i.e. when I was most likely to conceive.
- When it comes to /ʒ/, which contrasts with /ʃ/ in Czech (my mother tongue), I was surprised that here too I found menstrual phase to correlate with differences in the voicing properties of the consonant: during ovulation, there was the least amount of voicing. However, this was only supported by visual inspection and not the statistical analysis. Alas, trying to keep myself ignorant, little did I know that one token of /ʒ/ per day would leave me with only one token for ovulation per month… and thus only 8 tokens for the entire period of eight months.
- The voiced allophone of /ř/ showed the same effects as /ʒ/. However, perhaps because I had more tokens of this consonant for each day, this time the lowest amount of voicing associated with ovulation was also supported by the statistical analysis.
Just how much should we care though, really?
Although the potential role that menstrual hormonal changes may or may not have on language and speech is a topic that seems to stir a lot of interest, we should nevertheless bear the following in mind. Firstly, as far as I know, only three studies have looked into the potential effects of the menstrual cycle on linguistically meaningful phenomena (at least within the field of phonetics and phonology). Secondly, these studies suggest fairly small effects. Thirdly, we really don’t know just how widespread these small effects may be even within a single speech community, let alone in the world’s languages more generally. In light of this all, should we care about potential minute effects of the menstrual hormonal changes on consonantal variation? I would like to believe we should (otherwise I wouldn’t have invested so much energy in my method testing case study to begin with). Why?
The most immediate reason would be simply to reach a more complete understanding of all the factors that could possibly result in linguistic variation. Perhaps less immediately so, this would be related to two “problems” within the field of Language Variation and Change, as formulated by Weinreich et al. (1968): the constraints problem and the embedding problem. The constraints problem is related to what types of linguistic changes are possible. The embedding problem is concerned with what constraints linguistic variation.
The second reason to study potential effects of the menstrual cycle hormonal changes on linguistic variation is to understand the historical seeds of specific patterns found in language and speech, or indeed their evolution. For instance, if it is indeed the case that females have breathier voices than males cross-linguistically due to what’s known as glottal gaps – through which air can escape, see the picture above to the left again for my own posterior glottal gap – we may wonder why females should be more likely to display more glottal gaps than males to begin with. Could this ultimately boil down to the fact that females were more prone to breathier phonation than males on some days within their cycle due to hormonal changes? If that was indeed the case, such a feature may have been reanalysed for all sorts of socio-interactional purposes.
As yet, we don’t know that much on the role that menstrual hormonal changes may have on linguistic variation. As always in such cases, we need to boldly go where few have gone before.
Míša Hejná is an Assistant Professor at the Department of English at Aarhus University.
Abitbol, Jean, Jean de Brux, Ginette Millot, Marie-Françoise Masson, Odile Languille Mimoun, Helene Pau and Beatrice Abitbol. 1989. “Does a hormonal vocal cord cycle exist in women? Study of Vocal Premenstrual Syndrome in voice performers by videostroboscopy-glottography and cytology on 38 women”. Journal of Voice3(2): 157-162.
Amir, Ofer and Tal Biron-Shental. 2004. “The impact of hormonal fluctuations on female vocal folds”. Current Opinion in Otolaryngology and Head and Neck Surgery 12(3): 180-184.
Collins, Sarah A. and Caroline Missing. 2003. “Vocal and visual attractiveness are related in women”. Animal Behaviour 65: 997-1004.
Hall, Kira. 1995. “Lip service on fantasy lines”. In Kira Hall and Mary Bucholtz (eds.), Gender Articulated: Language and the Socially Constructed Self. NY: Routledge. 183-226.
Hall, Kira and Mary Bucholtz (eds.). 1995. Gender Articulated: Language and the Socially Constructed Self. NY: Routledge.
Hejná, Míša. 2019/Forthcoming. “A case study of menstrual cycle effects: global phonation or also local phonatory phenomena?”. Proceedings of the 19thInternational Congress of Phonetic Sciences, Melbourne.
Henton, C. G. and R. A. W. Bladon. 1985. “Breathiness in normal female speech: inefficiency versus desirability”. Language & Communication 5(3): 221-227.
Liu, Jinfeng, Dan Wang, Xiaoting Li and Wang Ningyu. 2017. “Association between sex and speech auditory brainstem responses in adults, and relationship to sex hormone levels”. Medical Science Monitor 23: 2275-2283.
Moisik, Scott. 2013. The Epilarynx in Speech. PhD thesis, University of Victoria.
Pipitone, R. Nathan and Gordon G. Gallup. 2008. “Women’s voice attractiveness varies across the menstrual cycle”.Evolution and Human Behavior 29: 268-274.
Raj, Anoop, Bulbul Gupta, Anindita Chowdhury and Shelly Chadha. 2008. “A study of voice changes in various phases of menstrual cycle and in postmenopausal women”. Journal of Voice 24(3): 363-368.
Wadnerkar, Meghana B., Patricia E. Cowell and Sandra P. Whiteside, S. P. 2006. “Speech across the menstrual cycle: a replication and extension study”. Neuroscience Letters 408: 21-24.
Weinreich, Uriel, William Labov, and Marvin I. Herzog. 1968. “Empirical foundations for a theory of language change”. In Directions for Historical Linguistics: A Symposium, edited by Winfred P. Lehmann and Yakov Malkiel, 97-195. Austin: University of Texas Press.
Whiteside, Sandra P., Anna Hanson and Patricia E. Cowell. 2004. “Hormones and temporal components of speech: sex differences and effects of menstrual cyclicity on speech”. Neuroscience Letters 367: 44–47.
*Or phonatory settings for those of us who are a bit nerdy, and also for those of us who prefer to be more accurate…
**You may find it difficult to believe that I was actually ignorant regarding how many phases one should work with. However, different studies work with different phases, and an initial dip in the literature left me more confused than I’d expected. See the references in Hejná (2019/Forthcoming).