Robert McColl Millar > Quotes

“the highly complex patterns of plural formation in Old English have been replaced almost entirely by the iconic -s plural, which has spread to many hundreds of words that formerly didn’t have it.”
― Robert McColl Millar, Trask's Historical Linguistics

“Another, much more famous, instance of morphologization has occurred in the Romance languages (the modern descendants of Latin). Latin had a noun mens ‘mind’, whose stem was ment- and whose ablative case-form was mente (the Latin ablative was a case-form with miscellaneous uses, most of which we would associate largely with prepositional use in English). Quite early, it became usual in Latin to use the ablative mente with an accompanying adjective to express the state of mind in which an action was performed; as was usual in Latin, the adjective had to agree with its noun mente as feminine singular ablative. We thus find phrases like devota mente ‘with a devout mind’ (i.e., ‘devoutly’) and clara mente ‘with a clear mind’ (i.e., ‘clear-headedly’). At this stage, however, the construction was possible only with adjectives denoting possible states of mind; other adjectives, like those meaning ‘new’ or ‘equal’ or ‘obvious’, could not appear with mente, because the result would have made no sense: something like ‘with an equal mind’ could hardly mean anything. But then speakers began to reinterpret the mente construction as describing not the state of mind of somebody doing something, but the manner in which it was done. Consequently, the construction was extended to a much larger range of adjectives, and new instances appeared, like lenta mente (lenta ‘slow’) and dulce mente (dulce ‘soft’), with the adjectives still in the appropriate grammatical form for agreement with the noun. As a result, the form mente was no longer regarded as a form of mens ‘mind’; it was taken instead as a purely grammatical marker expressing an adverbial function, and it was therefore reduced from a separate word to a suffix. Today this new suffix is the ordinary way of obtaining adverbs of manner in the Romance languages, entirely parallel to English -ly in slowly or carefully, and it can be added to almost any suitable adjective. Thus Spanish, for example, has igualmente ‘equally’ (igual ‘equal’) and absolutamente ‘absolutely’ (absoluta ‘absolute’). Spanish still retains a trace of the ancient pattern: when two such adverbs are conjoined, only the last takes the suffix, and hence Spaniards say lenta y seguramente ‘slowly and surely’, and not *lentamente y seguramente. In French, this is not possible, and a French-speaker must say lentement et sûrement.”
― Robert McColl Millar, Trask's Historical Linguistics

“For generations, linguists had very little idea what to make of this kind of variation. On the whole, most linguists were inclined to consider the speech of educated people as the primary object of description and investigation, while the vernacular speech of uneducated people was usually dismissed as being of no consequence – except in dialectology, in which the speech of elderly, uneducated, rural speakers was commonly considered to be the most suitable for investigation. Since earlier linguists were overwhelmingly male, there was perhaps also a comparable tendency to treat men’s speech as the norm, while women’s speech, where it differed, was often disregarded as inconsequential. Otherwise, however, the very high degree of variation within a single community was, for the most part, simply ignored: at best it was considered to be a peripheral and insignificant aspect of language, no more than erratic and even random departures from the norms, while at worst it was regarded as a considerable nuisance, as a collection of tiresome details getting in the way of good descriptions.”
― Robert McColl Millar, Trask's Historical Linguistics

“The progressive form of the verb go has been available for centuries in constructions such as I am going home, in which go clearly retains its ordinary verbal sense. This same construction could also be used with a complement of purpose, in cases like I am going to visit Mrs Pumphrey, in which the verb go still had its ordinary meaning: the structure of such a sentence was [I] [am going] [to visit Mrs Pumphrey], broadly parallel to [I] [am going] [home]. Such a sentence could be uttered by a speaker who was actually on her way to Mrs Pumphrey’s house, but equally, and crucially, it could be uttered by someone just about to set out, just like I’m going home. As a consequence, speakers began to reanalyse such utterances as expressing, not actual motion, but rather an intention for the near future. Accordingly, it became possible for something like I am going to buy a new carriage to be said by someone curled up comfortably at home with no immediate intention of moving. This largely happened in the early nineteenth century, but the new usage has extended its domain very rapidly, and today we routinely say things like You’re going to like this book, in which no relevant motion is even conceivable: the be going to construction has entirely lost its original connection with movement and become a mere grammatical marker of the (near) future. Together with this grammaticalization, the structure has been reanalysed: we no longer have the old structure [I] [am going] [to buy a new car]; instead, we have [I] [am] [going to] [buy a new car], in which going to forms part of a single grammatical marker. To see this, observe that this new going to can now be reduced to gonna, as in I’m gonna buy a new car. The same is not possible with the ordinary progressive of the verb go, as in *I’m gonna the beach, in which going and to do not constitute parts of a single grammatical form.”
― Robert McColl Millar, Trask's Historical Linguistics

“Earlier generations of European linguists, almost all of them speakers of IE languages, tended to assume without discussion that such a fine, upstanding proto-language as PIE, the speech of those splendid folk who carried their language into a huge area of the globe, must naturally have been equipped with a full set of numerals up to ‘100’, at least, and therefore tried stoutly to reconstruct a PIE numeral system of this size, invoking in the process any number of ‘replacements’ and ‘analogical formations’ to account for the wide discrepancies in the formation of the attested numerals. More recently, however, some specialists have begun to question this assumption, and to put forward the awful suggestion that the speakers of PIE could not count beyond ten. Their idea is that the numbers beyond ‘ten’ were created independently in the various daughter languages after they had diverged from the ancestral tongue. This makes a good deal of sense, and there is support for it. Most notably, the PIE word *kmtóm, found in all daughters and meaning ‘100’, did not necessarily have that precise sense in PIE. In Homeric Greek, the word seems to have meant simply ‘a large number’, while in Germanic it often means ‘120’ or ‘112’, as in the British hundredweight ‘112 pounds’. It therefore appears too rash to assume without discussion that PIE had numerals all the way up to 100 and even beyond. There is a moral here: even if a feature is found in all the daughter languages, we cannot presume it must necessarily have been present in the ancestral language, unless no other reasonable explanation is available.”
― Robert McColl Millar, Trask's Historical Linguistics

“The English adverbial suffix -ly has also been obtained by morphologization. Old English had a noun lic ‘body’, which has developed in various ways. As lich, it survives in lich-gate, a roofed gateway to a church where coffins were formerly placed to await the arrival of a clergyman. The derivative gelic ‘having a common body’ is the source of our word like, as in ‘She’s just like you’. But, early on, the word lic also came to be compounded with nouns to express the sense of ‘resembling’ and then ‘having the characteristics of’: hence Old English fœderlic ‘father-like’, ‘fatherly’ and manlic ‘man-like’, ‘manly’; here the original noun has since been reduced to a mere suffix. Finally, much the same thing happened with adjectives: a case-inflected form lice was added to an adjective to express the meaning ‘in the manner of’: hence Old English slawlice ‘slowly’ and cwiculice ‘quickly’, and here again the original noun has been reduced to a purely grammatical affix: our suffix -ly for making adverbs out of adjectives.”
― Robert McColl Millar, Trask's Historical Linguistics

“French provides a very striking case of multiple meta-analysis. Our word unicorn derives from Latin, in which it is composed of uni- ‘one’ and cornu ‘horn’. In English, nothing much has happened to this word, except that most speakers, knowing nothing of Latin, probably don’t assign any internal structure to it: they just regard it as a single morpheme, on a par with horse or giraffe. Most European languages have the identical word, but the French word is the curious licorne. Where did this come from? The original word, of course, was unicorne, a grammatically feminine noun. But the French word for ‘a’ with a feminine noun is une – and hence unicorne was misinterpreted as une icorne, and icorne therefore became the French name of the beast. But the French word for ‘the’ before a noun beginning with a vowel is l’. Hence ‘the unicorne’ was expressed as l’icorne – and this form in turn was reanalysed as a single noun licorne, producing the modern form.”
― Robert McColl Millar, Trask's Historical Linguistics