Saturday, 8 September 2018

Indo-European 'slippery'


I have been looking at the Indo-European family of words for the quality slippery, English slip and slide. I got into this because there is a Welsh word sglefrio, slide and I wondered if this had an English source. This is conformed by the standard dictionary. I found this rather disappointing. There is another Welsh word llithro, which is purely Celtic (from sglib-tro according to the reference work by Walde and Pokorny).
They list a substance as the primary element of meaning, so smooth mud, slime, etc., with slip as a verb following the presence of this substance.
Something surprising is that the same root describes smooth things and sharp things. These qualities are very different for an English speaker. But this does not show a fault in the proposed set of relationships. Schleifen in German means 'grind', both to remove roughness and make a surface smooth and to remove extra material and make a blade or point sharp. The action of removing tiny pattern-affecting burrs of material can make a given object either smooth or sharp.

The dictionary article for *lei (or *slei) starts with slime and goes on through Gaelic sleamhann, 'smooth', ὀλιβρόs 'slippery', λείμαξ 'snail', schleifen, schlűpfrig ('slippery)', Latin litura 'erasure'. There are other entries for “extensions” of this minimal root (which rarely appears without extensions). Latin lino ('smear') has a nasal infix like hundreds of other verbs, in the present stem only. The perfect and supine are livi litum (interesting variant levi) and here we see almost the naked root. Lei genuinely is the basic root and these forms display it.

In modern Gaelic, liofa means fluent in a language but is actually sharp, not literally 'fluent'. Liofa should be the past participle of a verb liomh or liobha, and in fact there is a verb liomh, which means 'sharpen'. Liofa is common in talking about the Gaelic language, where the shrinking number of fluent speakers is a constant cause of concern. This is a distinctive assignment of symbolism in Gaelic and perhaps in Welsh. (The 1927 dictionary spells this liobha, which is pronounced the same way. Liobha is related to schleifen and presumably Latin limare.)

O'Huigin describes a word in Old Irish rind, which means blade but also mouth. The idea is that language is sharp and the mouth as the location of language is therefore like a blade. In Welsh, min means 'lip' but miniog means 'sharp'. This may not be a phonetic coincidence, it may be the same metaphor as in Old Irish.
Another word for sharp has an important range of applications to language: geur. Geur-cainnte is used to describe bardic ability, in a famous historical description of the situation in Scotland. Sharpness is the metaphor for a highly desired quality of language. It is not as simple as hostility, it means acuity as well.

Looking at 100 cognates within the posited *lei root shows a beautiful pattern. The strength of the IE idea is reinforced as one looks at the overall similarities across many languages. Quite clearly this pattern becomes simpler as we go back in time. But the point of origin of the pattern looks increasingly blurred. It is like an old photograph which is mainly damage to the substrate – within which real information is hidden. This follows directly from the path that was followed – the posited “Ursprache” is the last stage before knowledge vanishes completely, so inevitably part of the Ursprache is not known, a blur.
Walde-Pokorny defines the root as *lei and suggests that the real words we know about are built up by suffixes, variously -b(h) (slip), -dh (slide), -m (sleamhann), -t (litura). Besides this, most derivatives also add an s- as an affix; so that slide corresponds to Latin lubricus. The most naked form of the root shows in Latin lino and litura.
If I am correct, the reconstruction of the “*lei” root involves five different root suffixes. It is hard to avoid the impression that these are being made up ad hoc to account for original inconsistencies within the posited language. That would mean that the reconstruction method had failed – and perhaps that its object frustrated logic because it is irregular and disunited. We are applying a perfectly logical method to a subject which does not answer consistently; perhaps the efforts of scholarship have shown that Indo-European is not a solid object – perhaps a dialect continuum, perhaps a language in which different patterns were co-present with different frequencies. So the language of 3000 BC was complex – much of it was lost – what survives is simpler than the original – but has diversified while spreading through Europe to produce a new, secondary complexity.

I could not find the Russian word skolzkiy, (or verb skol'znut') in W-P, surely related even if that shows an irregular process of development. The real pattern may be even more irregular than they allow for. The root appears to include a -g- which gives rise to the k in skolz and the g in glide. (And the k in Norwegian sklie.)

We don't have any IE words, so the roots are the substance of our patiently acquired knowledge. It would be better if the roots were more clearly demarcated, or if the extensions which turn them into words were better understood or more regular.
There is a core problem which surfaces in the W-P entries for roots connected to slip. They list roughly ten separate roots but then admit, in remarks within the article, that some of these roots are extensions of other ones. So it comes back down to one or to two or three, in a frustrating way. The problem is that these variants do not fit together by sound processes which we can sustain by analogies, and also that they are too close to each other to be totally separate. The gaps can be made up by positing root extensions, but this has the fault of being arbitrary. For an English speaker, the idea that “slip” and “slide” are really the same word, is attractive and natural; but is there any other pair of words that shows the same alternation? The reality of the imagined suffixes depends on the frequency with which they show up; otherwise they are just invented solutions to a problem which has not actually been solved. The material is at once too rich, an excess of related forms; and too poor, as the relationships are not validated – even though there are so many IE relationships which we can validate, so many times. Walde-Pokorny does not admit that slip and slide relate to the same root. It seems likely that they are from one root, and that this is blurred or branching within Indo-European, and also that the words slick, sleek (and cognates) are yet another variant on the same root. Obviously, a slick surface is one on which you are likely to slip. Why don't these forms converge? Is the convergence we do accept partly the product of rules invented by scholars? What does half-convergence mean?

The three-letter (trilitère) theory about the rules of the Indo-European root eliminate *lei as a possibility. However, the consensus is that Ce roots do exist and the triliteral ideal is just a norm. Walde-Pokorny dates from before laryngeal theory was the standard model and takes no account of laryngeals at all.
Struggling through the examples reveals that there are almost none from Indo-Iranian, although quite a few from Slavonic. This gives another way of testing the kinship between the ten or so roots which W-P separate, so slip and slide. If they follow the same geography, it is more likely that they are indeed the same root. Without doing the number-crunching, I nonetheless suggest that they have very similar geography and are European as opposed to something older. One way in which a word pops up with many very similar forms, which are hard to relate to each other, is that it is a loanword from another language in which phonology and word-formation are strange from the point of view of the accepting language.

The big Welsh dictionary gives sglefrio as a derivative from English sclither (for slither). Sclither is a Scots form and I don't think it exists in the parts near Wales. So, a problem. I believe that sl- is Anglo-Saxon and the group of forms in scl- are Scandinavian loan-words. The OED gives sklither as a obsolete form of slither. I have much used a book called Sglefrio ar eiriau, 'slipping on words', i.e. getting words wrong, a book about Welsh poetry not written by Methodist clergymen.
I can't resist citing two Scots dialect words omitted by W-P. This is in the context of the basic element of the root being 'soft mud' and a collateral form in sleug-, sleuk. Scots has sleek 'mire, slime; miry clay in the bed of a river, the seashore'. Further, sleech 'silt; sea-wrack; the oozy, vegetable substance found in river-beds; slime; in pl. foreshores on which silt is deposited by the tide, 'slob-lands'.' OED says that slob comes from Irish slab, mud. English cognates of sleech would be slough and slack, a dialect word meaning swamp or soft ground (Gordale Slack and so on). We read that “much of central Belfast is underlain [...] by a deposit of soft grey mud, silt and fine sand with numerous sea shells, in particular oysters” known as Belfast Sleech. It sounds as if slob-lands have matured into high-quality sleech.

I read about these stray consonants in Szemerenyi's standard work of 1990. Szemerenyi says that the best work on them is still by Per Persson (1857-1929). I found Persson's book of 1912 – a thousand pages long, of which half is about the root-determinatives. He had published another book on the same subject in 1891, and the 1912 one adds a crushing level of detail and retorts to his critics. It looks as if Persson's tenacity and command of facts had killed the question – a century later, one can conclude that nobody was able to improve on him, and wonder how many people have read the book in the intervening century.
The issue is well summed up by James Clackson. He compares Greek kheo with Latin fundo and German giessen (noun form Guss) and points out that these look like the same root except that the expected d is missing in Greek. The d is called a Wurzeldeterminativ, or root extension. (German also has schütten, the same root with an s- prefix which we see in hundreds of roots.) We have 3 options here:
(a) kheo is not related to fundo
(b) the Greek word is damaged in some way (perhaps by vandals)
(c) the -d is a separate element which (like the s-) was attached in some cases but not all and whose meaning we do not know

Obviously, these root extensions have a bearing on the relationship between slip, slide, and slick. Persson piles up hundreds of roots in which these extensions sometimes appear. He says that virtually every consonant can appear in this role – which suggests that they are not elements of specific meaning (as -n- generally makes a root into a present stem) but more generalised. Thus, slip slide and slick are not differentiated by meaning (and the possible -p -d -k extensions are not signs which restrict or refine the meaning). Persson says that there is no known explanation for the extensions – they are simply there, like the endings for noun cases or persons of the verb. He believes they once had meanings. He does not comment at all on geography. Apparently all parts of the Indo-European world use the extensions in the same way. (Persson wrote before Tocharian or the Anatolian languages had been uncovered.) This could suggest that the phenomenon is very old, in fact primordial. The variation between slip slide slick looks like a geographical variation, but all the variants appear in English, which shows the contrary. Hypothetically, the source of Germanic could be a migratory war-band which had drawn recruits from different parts of the Indo-European dialect area, so that geographical variants all became established within one new speech community (which then separated from the source population). This would be like the captain:chieftain dualism, where a dialect variation in Old French produced two traces in English. There is no evidence for this, though. More geographical analysis is possible, but Persson did not stumble across any geographical patterns in his exhaustive work. This is a basic problem with any idea based on geographical variation. (Strike war-band, insert “migrant labourers in the sheep industry”.)
The unmoved validity of root-extensions makes it difficult to count Indo-European roots. What appear as ten articles in Walde-Pokorny could represent one root with interestingly oscillating extensions.

The s-mobile also appears in a root *ker, variant *sker-. This shows up in English as shear (as in shear through). This is obviously (obviously?) related to sharp. The -p is unexplained and so looks like a root extension.  

No comments:

Post a Comment