Tracing the linguistic sound-shifts of �h� and �g�

L M Leteane | Friday July 22, 2016 13:00

For Sotho-Tswana speakers, the most striking difference in pronunciation as regards these two consonants is between Sotho and Tswana. Sesotho has a partiality towards h for certain terms where Setswana prefers g (a guttural one, pronounced like the J in Julio), e.g. hare (middle) instead of gare, holo (great, large) instead of (k)golo. In ancient Egyptian, I have noticed, this same h is often pronounced like the Setswana g, and yet obtains in words that, today, we would use r for. An example of this is ‘Tehuti’ (‘Thoth’), the name of a famous Egyptian Wisdom-god. The ‘huti’ is actually ruti in today’s parlance – suggesting that it was pronounced as the French still pronounce r (which is something like the guttural Setswana g).

Other h-to-r sound-shifts that mutated between language families include hiri (hair) – which became riri in Setswana, and the phrase pele ho (‘besides’: in Sotho pele is not just ‘in front’, it can mean ‘besides’), thus the Latin term pleuro (‘next to’). But there are instances where the r in Setswana corresponds well to the Indo-European r for given proto-terms. Examples include rulela (‘put on top’) and ‘rule’ (boss over, be on top); thari and ‘tarry’ (both meaning ‘[be] late, behind time’). The most consistent instance of this is when the r is onomatopoeic (echoes the natural sound obtaining in nature), thus thoromo and ‘tremor’, rora and ‘roar’. In similar vein, when formerly orderly components fall apart (and one imagines a rumbling sound as this happens), we get kgoropa (break down) and ‘corrupt’, also meaning ‘break down’.

There are also instances where the h is enunciated the same in Bantu and Indo-European languages. Examples include hula, and ‘haul’, but in Setswana hula can also be fula. Hika (Greek) and heka (’mount, be on top of’ – now used mainly in a sexual context) and its derivative hekeetsa (‘overwhelm’). There is also le-hala and ‘hull’ (‘empty space’) and its derivatives se-hala and ‘shell’. So, as we can see, se + h can sound-shift to sh in both Bantu and European, just as we saw last week when se+ huba (‘hub-like’) became shaba (‘share, trade’) – a hub-like activity. 

There are also cases where the consonants h and r are enunciated in certain language families – but not in others – for the same proto-terms. Thus, for h we have bohola, ‘bark [at]’) and ‘bawl’ (shout at, cry loudly). In the case of r we have khutshwana and kurtzwile (German), both meaning ‘short’, and their variations hutsa and ‘curse’ (both meaning ‘make (or wish) one not to grow (or succeed)’, thus make another ‘stunted’ or ‘short’). In similar vein, the English would write the r, but not pronounce it, for the term ‘curse’ – whereas the Americans (and Irish) would enunciate the r. However, from a purely ‘proto-term’ perspective, we find that, primordially, the now-absent r in khutshwana was indeed enunciated long back in time because kuru means ‘curved’ (as in le-koro, and koro-bala)…a shape or posture that makes one appear shorter than should otherwise be the case…and even Setswana inadvertently acknowledges this.

There are also many cases whereby the r sound-shifts to l, and we have already seen plenty of examples of that in this column.  But, by way of reinforcing that fact, we can look at more examples. An interesting one is se-bola (‘stripe-like’) which we can relate to ‘zebra’. Evidently, we had this term long ago, but due to time and circumstance it was lost and the rather bland pitse-ya-naga (literally ‘wild horse’) instead took over. I cannot help thinking that if the l to r sound-shift was long better-discerned here, we might still be calling it pitse-sebola! Other examples include the Latin term quare (‘old, ancient’) and the Setswana term kgale; the Latin term pre (‘of before’) and the Setswana term pele. There is also the Greek term aphaire (‘to snatch [from]’) which equates to the Setswana term a phaila (‘he snatched’). In similar vein we have tolang-se – which evidently sound-shifted, in Latin, to ‘trans’…but both meaning ‘[to go] across [something]’. An extreme case, if you like…in the sense that it incorporates both the h-to-r shift and the l-to-r shift…is found in a derivative of the term hula. This, I suggest, is how hula became first gola (‘draw, thus pull, [payment]’, and gola (different intonation): ‘grow’, thus ‘extend, be pulled out’), before it became rura, the most r-centred case, which pronounced as goga (pull).

 Now, the hula (pull) we noted above – which is also ‘graze’ in Setswana because, when grazing, an animal pulls out grass from the ground – can also be enunciated as fula. Apparently, this h-to-f shift exists in Indo-European as well because fulaka (‘graze widely’) relates well to ‘forage’ – albeit with a discernible l to r sound-shift, as well as the tendency for g (as in ‘goat’) to be hardened to a k in Setswana: for example, in Setswana, ‘Kalanga’ is simply ‘Kalaka’. Another example of the h-to-f shift across language families is the term ‘conifer’. Its base-word is ‘cone’. Now, a cone tapers as if a parallel shape was ‘hammered and bent’ into a sharp point (hence a ‘coniferous tree’ – e.g. a fir tree – is one that exhibits a cone–like or tapering shape). As such, ‘conifer’ relates well to the Setswana term koneha (bend). So, although koneha is now enunciated as konega in Setswana – i.e. with a guttural g – we can nevertheless see that the h-to-f sound-shift is not confined to Bantu languages only, but that it can spill over to other language families for the same proto-term.

Comments to leteanelm@gmail.com