Tone and Humidity.

A startling theory (description from the FAQ):

Everett, Blasi & Roberts (2015) review literature on how inhaling dry air affects phonation, suggesting that lexical tone is harder to produce and perceive in dry environments. This leads to a prediction that languages should adapt to this pressure, so that lexical tone should not be found in dry climates, and the paper presents statistical evidence in favour of this prediction.

It’s reminiscent of this bit of japery, but this is serious, and the FAQ answers some obvious questions, e.g., re exceptions:

There are certainly exceptions to the prediction. But we’re not expecting a hard and fast rule, just a statistical tendency. We were aware of such exceptions before even embarking on the analysis of the database. Our account is not simply deterministic, but suggestive of gradual pressures operating at the same time as other pressures known to impact the evolution of sound systems. Occasionally such influences may even be at odds. We note in the SI, however, that even in a language family like Tibeto-Burman, in which there are exceptions, the overall pattern holds in the predicted direction. Tibeto-Burman languages spoken in more desiccated regions are less likely to employ complex tone.

Thanks, Yoram!


  1. Just an FYI: Caleb Everett is Dan Everett’s (he of the Pirahã fame) son.

  2. I have not read the article and I would not be qualified to offer an informed opinion. But it seems plausible to me that sound would travel differently in dry and humid air. And, that this would have some effect on perception of sounds. And, over time, this could effect the phonology of languages.

  3. Sure, it seems plausible to me too. But I’m just some patzer, and we all know what plausibility is worth.

  4. Trond Engen says

    Doesn’t seem plausible to me, but I’m definitely nobody.

    If the mechanism is real, I’d expect tonal languages spoken in dry climates to have greater difference between tones than equally complex tonal languages spoken in wet climates.

  5. It’s easy to knock this down with counter-examples but the real test is if they used a valid sample to analyze. The Apachean languages are all tonal, and in the case of Navajo there are functionally four tones. It can get dryer than the Southwest, but not by much. Speaking of four tones, Mandarin has four and northern China is a dust bowl in the winter. Ditto for Tibetan. Where you find the best sample is in Africa, where you have tonal languages in areas ranging from coastal forests to Sahelian savannahs. The same is true of east Africa.

    Maybe you get more tones in wetter climates – Cantonese versus Mandarin – but Khmer has none at all. the Hmongic languages, at higher elevations, are all tonal. In fact they top everyone when it comes to having tones.

  6. The study focused on complex tones as an indicator for the effect they claim, so Navajo and many Bantu languages are binned with the not-so-tonal languages. The authors say that using tonal vs. non-tonal gives similar results.

    I wonder, do speakers of Thai, Cantonese or Trique who live in Phoenix or Riverside have a hard time understanding each other? Do people who sing outdoors in L.A., Madrid or Jaipur find it extra difficult to produce clear tones during the summer? Someone must have noted something by now.

  7. Jongseong Park says

    Trond Engen: If the mechanism is real, I’d expect tonal languages spoken in dry climates to have greater difference between tones than equally complex tonal languages spoken in wet climates.

    Their thesis seems to be that lexical tone is harder both to perceive and to produce in drier climates, which I would guess to mean that producing greater difference between tones would also be harder.

    One way to enrich the data could be to investigate dialects that belong to the same language, like Mandarin, whose dialects are spoken over a wide swathe of China with a significant range of climates. But I don’t think Mandarin dialects display significant variation in the complexity of the tonal systems, with most having four tones just like Standard Mandarin.

    I’m all for taking a crack at investigating those numerous claims about the influence of climate on language, such as the influence of humidity on different kinds of phonation and classes of sounds like fricatives. We hear these old nuggets like the Australian accent of English arising from having to speak with the mouth closed to keep out the dust. We’re not meant to take these claims seriously, but imagine we actually investigated how various environmental factors affect openness and tongue root position in the vowel inventory. We’re probably not going to find anything conclusive, but it would be a fun exercise.

  8. Trond Engen says

    Differences would be harder to percieve, so greater effort would have to be made for the same level of tonal complexity to be transmitted. So a language magically brought into a dryer environment might either reduce transmitted complexity or increase the effort. Either way, for any given level of tonal complexity, more articulatory effort should be involved.

  9. Just to go back to Everett et al., the claim for the mechanism of dry air impacting tone production comes from a paper by Hemler et al., The effect of relative humidity of inhaled air on acoustic parameters of voice in normal subjects. I have only seen the abstract. What Hemler et al. find is that test subjects, saying a prolonged “aaa…” in dry air, have more jitter in their voice compared to doing the same in normal and humid air. Presumably that is due to dehydration of the surface of the vocal cords. By inference, more jitter (fluctuation in pitch) should make tone harder to perceive.

    One may ask, how dry is “dry air” in the experiment? Does the effect kick in at 10% humidity? 2%? 30%? Is the jitter noticeable over the duration of a single vowel? Is the pitch variation significant compared to that due to tone? Is it perceptible by human ears? Do speakers of tonal languages have trouble perceiving lexical tone of speakers with high jitter from other causes, like age?

  10. Oh for God’s sake, already.

    Dediu: Ultraviolet light affects the color vocabulary.

    It has been suggested that people living in regions with a high incidence of ultraviolet light, particularly in the B band (UV-B), suffer a phototoxic effect during their lifetime. This effect, known as lens brunescence, negatively impacts the perception of visible light in the “blue” part of the spectrum, which, in turn, reduces the probability that the lexicon of languages spoken in such regions contains a word specifically denoting “blue.”

    Ed. See previous discussions in Nature (Open Access), here, here, and here.

  11. I would imagine that the effect, as Y says, would have more to do with the drying of the vocal chords by inhaling dry air than by changes to the propagation of sound while exhaling, since exhaled breath is (almost) always very wet.

  12. They will be the death of me yet. Wang et al., Temperature shapes language sonority: Revalidation from a large dataset. PNAS Nexus 2(12), 12/2023 (Open Access):

    Multiple factors of the natural environment have been found to impact and mold the phonetic patterns of human speech, among which the potential correlation between sonority and temperature has garnered significant attention. We leverage a large database containing basic vocabularies of 5,293 languages and calculate the average sonority for each language by adopting a universal sonority scale. Our findings confirm a positive correlation between sonority and temperature across macroareas and language families, whereas this relationship cannot be discerned within language families. We suggest that the adaptation of the distribution of speech sounds within languages is a slow process which is moreover insensitive to minor differences in temperature experienced by speakers as they carry their languages to new regions. Nevertheless, at the global level a solid relationship emerges. Furthermore, we delve deeper into the nature of the relationship and contend that it is mainly due to cold temperatures having a weakening effect on sonority. This research provides compelling additional evidence that climatic factors contribute to shaping language and its evolution.

    Popular account here (

  13. Was Jacobus van Ginneken (1877-1945) not the first linguist to posit the influence of climate on language?

  14. says:

    “In these languages, words take on entirely different meanings depending on the lexical tones in which they are spoken,” Deutsch said. (Lexical tones are the pitches of words.) If you say the word “ma” in Mandarin in one tone, for instance, it means “mother,” but in another tone, it can mean “horse,” she said.

    Does this interfere with singing?

  15. Does this interfere with singing?

    In the case of Chinese languages, yes.

    Canto-pop has a bunch of conventions (that wikilink doesn’t really explain perhaps this) whereby songs keep ‘in tune’ but don’t mangle the sense too bad.

    Classical Chinese Opera also incorporates work-arounds. With again Cantonese style being the most highly-developed.

    Most Chinese ‘words’ can be pronounced with ‘neutral’ tone in unaccented contexts/’where the meaning is clear’, so there’s increased emphasis on articulating the consonants. (OTOH, it seems to me that audiences don’t pay a lot of attention to the text/words so much as to the dramatic action.)

  16. @David W, by the way, that article seems a lot confused

    different meanings depending on the lexical tones in which they are spoken

    Tonal languages usually distinguish by ‘relative tone’ — which the article also discusses. (And commenters here are far better qualified than me to elaborate.) Then absolute pitch has nothing to do with it.

    If a language relied on absolute pitch, kids/women with higher voices (because smaller stature) would struggle to get the lower pitches, and vice versa.

    difficult to tease apart the influence of genetics from a person’s environment, for instance parents with absolute pitch may be more likely to spend time teaching their children.

    In Western cultures, it is musicians more likely to have absolute pitch. And musicians as parents are more likely to give musical training/awareness to their kids. [Check out Rick Beato and his son on Youtube, for example. Rick has ‘lost’ his absolute pitch in recent years, a mixed blessing. His son’s is (so far) very sensitive.]

    _If_ a Chinese-speaking upbringing is more likely to train absolute pitch, it could well be from cramming music into the kids. Nothing to do with the language.

  17. See also Maggu et al., Combination of absolute pitch and tone language experience enhances lexical tone perception (here).

  18. Thanks @Y

    Absolute pitch (AP), … has prevalence estimates ranging from 0.01 to 1%

    And wikip has an even vaguer range. So _if_ there’s a greater prevalence amongst East Asians (evidence doubtful, and not supported for those of E.Asian ethnicity growing up in Western countries), AP is hardly a requirement to speak Chinese/tonal languages.

    I do wonder whether the (alleged) findings are statistically significant. The studies I’ve seen are on tiny samples. (And I’d have a suspicion those were drawn from populations easily accessible for the researchers. So Linguistics and Music students.)

  19. Modern Mandarin pop singing, however, does not match tone to the melody (or vice versa). There’s enough redundancy in the segmental phonemes to carry most of the meaning most of the time. No doubt this is because there are so few lexical tonemes. Indeed, some dialects of Mandarin have only three tones or, like Dungan, merge tones 1 and 2 in the non-final syllables of phonetic words.

  20. David Eddyshaw says

    I know very few Kusaal pop songs, but they seem to ignore tone. Tone doesn’t carry a great lexical load in Kusaal, and you can generally second-guess syntactically-determined tonal changes from context, so problems of actual comprehension don’t really arise. People who can read Kusaal manage fine without any tone marking, not even the occasional disambiguating one that is used in a lot of the standard orthographies in West Africa.

    Moreover, Kusaal full words carry stress on the root syllable, and tone sandhi works in such a way that except after pause, the great majority of full words carry initial mid or high tone, with low tone mostly appearing on clitics, so there isn’t a lot of noticeable clashing between tone and stress in connected speech.

    I wonder if anyone’s looked at Yoruba in this regard? There’s a lot of pop music in Yoruba.
    There’s been some work on tone in more traditional Yoruba genres, anyhow:

  21. (OTOH, it seems to me that audiences don’t pay a lot of attention to the text/words so much as to the dramatic action.)
    Basically the same as with Classical European opera? Without consulting a written text, I’m normally not able to understand what the singers sing in most arias, as it is so distorted, and I assume I’m not the only one.

  22. Basically the same as with Classical European opera?

    Hmm, which mostly goes on in non-English, so I’d have a hard time even if the words were clear. I get indigestion with the really ornate stuff anyway — prefer Benjamin Britten or Prokofiev.

    Classical Chinese Opera mixes full-on warbling with recitative with dialogue. (I guess the nearest parallels would be Baroque Oratorios/Passions.) So there’s plenty of opportunity to keep up with the plot. (Not that the plots seem anything but ridiculously contrived — so we’re back to Classical European again.)

  23. There are different styles of classical European opera singing, and some singers have diction* that is quite easy to understand. Listen to this, which contrasts the quintessentially operatic tenor singing of Placido Domingo with John Denver’s folk tenor delivery. Despite Domingo’s strong accent and stylized delivery, I have never had any trouble understanding what he was saying—and my hearing for lyrics is generally pretty poor. Here is another one, which compares Maria Callas and Renata Tebaldi singing the “Habanera” from Carmen. I find Tebaldi’s lirico-spinto delivery much easier to make out than (and superior to) Callas’s bel canto** style.

    * Why is this sense of diction almost solely limited to discussions of singing? We don’t normally use the word to describe clarity of enunciation for non-musical speech.

    ** There is a lot of debate about what bel canto should really mean, but whatever it is exactly, everyone seems to agree that Callas’s operatic style was one of the exemplars of it in the twentieth century.

  24. Do any of the “climate affects tone” articles discuss Hindi and Panjabi? One tonal, one not, yet spoken not only in the same climate but very, very often in the same location (Panjabi’s influence on Hindi in Delhi, the home of Khadi Boli, is of such longstanding “infamy” Snell uses in in his Teach Yourself Hindi series). Back when I could read Gurmukhi a Panjabi friend told me “you read Panjabi like it was Hindi”, testament to the to the impact of its tomes and my inability to grasp them.

    Admittedly, Panjabi only has two tones, but I’d be interested to know if the two are discussed, or if Panjabi is lumped in the “we know there are exceptions” category.

  25. Why is this sense of diction almost solely limited to discussions of singing? We don’t normally use the word to describe clarity of enunciation for non-musical speech.

    Actors, public speakers, and preachers, too.

  26. @Robin: AFAIK there was just the one article, by Everett et al. I don’t know if it includes Hindi and Panjabi (the database they utilized is offline), but I would guess both would count as negative, since they were only looking at “complex tone”. Anyway there are exceptions galore everywhere, despite the weak statistical correlation the authors claim.

  27. Y, thanks! 🙂

  28. David Eddyshaw says

    The original paper’s assertion that Niger-Congo languages without tone are much more likely to be found in arid regions may alarm the speakers of Atlantic languages, accustomed as they are to rice as their staple diet. No doubt drought is a major problem on the Kenya coast, too. I well remember the deserts of Mombasa.

    Frankly, anybody who could write such extraordinary piffle and not wonder if something has gone badly wrong with their methodology is evidently not one to let a brutal fact slay a beautiful theory.

  29. David Marjanović says

    Canto-pop has a bunch of conventions (that wikilink doesn’t really explain perhaps this) whereby songs keep ‘in tune’ but don’t mangle the sense too bad.

    Traditionally, Cantonese does it the other way around: 1) exaggerate the range of your speaking voice so the 3 level and the 3 contour tones become distinguishable as 6 level tones; 2) read the lyrics aloud that way – and you’re singing. The tones are the tune.

Speak Your Mind