What is the explanation behind the /p/ to /h/ phonological change from Halegannada to Kannada?

  • 1
    I have added more tags. Are the "closers" happy now?
    – fdb
    Commented Nov 23, 2018 at 10:37
  • Wikipedia doesn't indicate Kannada as having a /h/ phoneme. Can you please edit this to explain more.
    – curiousdannii
    Commented Nov 24, 2018 at 10:57
  • @curiousdannii: Kannada ಹ is /h/, not /ħ/. Listen to the recording here: omniglot.com/writing/kannada.htm
    – fdb
    Commented Nov 24, 2018 at 16:55
  • 1
    @curiousdannii The German Wikipedia article says it is /h/ (it is written independently of the English version) de.wikipedia.org/wiki/Kannada Commented Nov 24, 2018 at 22:49
  • 2
    @curiousdannii. If there are no minimal pairs for [h] versus [ħ] then you can in any case consider them as allophones of the phoneme /h/.
    – fdb
    Commented Nov 25, 2018 at 11:54

5 Answers 5


Kannada is not unique in this. Indo-European *p becomes h in Armenian, as in hair "father".

  • 3
    Also, Indo-European *p disappeared in Celtic, and (at least according to some authorities) early Japanese *p became /h/.
    – Colin Fine
    Commented Nov 23, 2018 at 10:52

As @fdb already noted, this sound change is not rare cross-linguistically. It is typically not a one step process but a chain of sound shifts /p/ -> /pʰ/ -> /f/ or /ɸ/ -> /h/ (and finally /h/ -> nothing; as observed in the evolution of the Celtic languages from Proto-Indogermanic).

All the steps are frequently attested in isolation, e.g., /p/ -> /f/ from Proto-Semitic to Arabic, or from Proto-Indogermanic to Proto-Germanic; /f/ -> /h/ for Spanish.


Before listing possible reasons for the change, below mentioned facts should be considered:

1) We can't pronounce words with /p/ much louder compare to /h/ (Try shouting Pogu and Hogu and observe the difference).

2) Not all the /p/ to /h/ phonological change occurred in Kannada. -Rarely used words are not changed (ex: Pashana(means poison)). -Words used in literatures are not changed (ex: Panditya, Prakara, Prashamse).

3) Labour class people using common words are mainly changed (Haru, Hattu).

4) Other Dravidian language still retain /p/ instead of /h/ changes. -Tulu, Tamil and Malayalam still retain /p/ in same words (ex: Pola/Po instead of Hogu).

Possible Reasons:

Geographical terrain plays major role in change behind /p/ to /h/

-Kannada speaking area is called Bayalu Seeme which means plain ground, this area is surrounded by Western ghats and Eastern ghats and hence difficult to maintain communication with surrounding Tamil, Malayalam and Tamil people.

-In large agricultural ground area need to speak in louder with someone in longer distance(Now also construction workers from Bayalu seeme speak loudely).

  • 2
    I think Adam Farris was asking about phonetic (articulatory or acoustic) mechanisms, Karthik. All language change has its distribution affected by geographical factors, but it's difficult to sustain an argument that a particular change did or didn't happen because of them. It's also difficult to support the argument that a particular change happened because people needed to speak louder. And I'm guessing that the literary words you refer to are mostly borrowed from Sanskrit, and borrowings often do not exhibit the sound changes of the rest of the language.
    – Colin Fine
    Commented Jan 3, 2019 at 18:29

This phenomenon is called debuccalization. Generally, this phonetic mutation is contextual. For example, in Berber (see Chenwi language) or Germanic (e.g.: hundred), it occurs in word-initial position. These debuccalized sounds had generally a primary or secondary posterior articulatory. So, maybe, your 'p' is not just a [p], but an aspirated consonant.

  • 2
    Germanic languages didn't shift /p/ -> /h/. German hundred is cognate to Latin centum, there is a regular sound correspondence German /h/ ~ Latin /k/. Commented Nov 23, 2018 at 16:25
  • 1
    Where do I said that Germanic /h/ shifted from *p? I am talking about debuccalization. Any sound that became a pharyngeal or glottal consonant. For example, in some Berber languages, initial [t] is pronounced [h] (tamttut > hamttut = woman). I am trying to widen this subject, not just focusing on p > h.
    – amegnunsen
    Commented Nov 23, 2018 at 16:38
  • 3
    I just wanted to be crystal clear here because the original question was specifically about the shift /p/ -> /h/. Commented Nov 23, 2018 at 16:41
  • @jknappen. Exactly.
    – fdb
    Commented Nov 23, 2018 at 17:47

This /p/ to /h/ transformation is indeed quite intriguing.

The theory about /p/ getting aspirated and eventually losing the labial won't work because that could have happened to even /b/ or /t/. Further, the already aspirated labial words imported from Sanskrit, like "bharata" (native name for India) and "phala" (fruit) always lose the aspiration rather than the labials in popular speech. The theory is quite academic and highly flawed.

If anything the /k/ should have become /h/ as it is quite close, and rapid or careless speech can easily render that sound. And guess what, it indeed happens, but in Tamil. To Kannada speakers its quite funny because not only the rule is not consistent, its also inverted. Eg. "mohan" is said and written as "mogan" but "pustakam" is said "puttaham".

The main problem in tracing the transformation is that the documented language Halegannada is highly conservative. It persisted even in 13th century when the 12th century Lingayat movement already produced works in spoken register, at least of the northern regions, called Nadugannada (middle Kannada) and already has features of modern Kannada (Hosagannada), including words with initial /h/. Eg: the most famous vachana of Basavanna ಉಳ್ಳವರು ಶಿವಾಲಯವ ಮಾಡುವರು uses the adjective ಹೊನ್ನ (honna) and not ಪೊನ್ನ (ponna). That said, the definitive grammatical work Shabdamanidarpana which was the basis for future works was written in mid 13th century in Halagannada but makes no mention of the transformation. This means it is very hard to determine when and how exactly did the transformation happen, because it suggests that spoken register was widely different from conservative written register and no one has recorded any intermediate aspirated forms like ಫೋನ್ನ (phonna) for example.

To speculate alternate theory where initial /p/ was lost leading to pogu -> ogu -> hogu, such intermediate forms are also not documented anywhere, not even in stone inscriptions that dot the entire Kannada speaking regions, much beyond modern day Karnataka state, which were basically meant for general public who are not necessarily well versed in literary register. While there are several idiosyncrasies wrt spelling the p-h transformation is remarkably stable, either being /p/ or /h/.

One theory suggests that it is due to the direct influence from Prakrit, and was adopted along with the /v/ /b/ transformation that is also unique to Kannada among Dravidian languages. However this is also flawed because the conservative Halagannada always uses /p/ along with initial /b/ in place of /v/; native words with initial /v/ transformed to /b/ very early and Sanskrit imports (tadbhava) already show /b/ /v/ transformation at initial position while other positions are inconsistent. Further, the custodians of the most ancient Kannada literature available were Jains whose sacred/liturgical language was Prakrit not Sanskrit, but even they maintained the /p/, but used /b/ abundantly suggesting that the /b/ was already firmly in place.

The Prakrit theory seems sensible but it has at least two problems. First, why did /p/ /h/ take longer than /v/ /b/? Second, Prakrit has dozens of morphological characteristics that transforms it from Sanskrit yet only two of those seem to be applied to Kannada, why? Further, Jainism was never a majority religion; popular religion fluctuated widely across generations due to changing royal patronage, so while we know definitely that overwhelming majority of the earliest poets were Jains, later to be joined by Lingayats, we have no clue about the proportion in general population, thereby making the Prakrit influence somewhat questionable since non Jains don't regard it as liturgical or important.

The intrigue continues.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.