Re: ubuppa

Late Latin _ubuppa_ 'feeding bottle' is attested once in Mustio's translation of Soranus's Gynaecia (V. Rose, Sorani Gynaeciorum vetus translatio latina, Lipsiae 1882, p. 43):

"131. Quid ei bibere dabimus? Aliquando aquam aliquando vinum aquatius per vasculum vitreum ad similitudinem papillae formatum et pertusum, quod rustici ubuppam appellant aut titinam."

"What shall we give him [the infant being weaned] to drink? Sometimes water, sometimes wine rather watered down, by means of a glass vessel made into the shape of a teat and bored through, which country folks call 'ubuppa' or 'titina'."

I have not found a satisfactory published etymology of _ubuppa_. The fact that Mustio (or Muscio, or Moschio) was from North Africa does not mean that the word originated there. M.'s task was to make Soranus's medical works accessible to women who spoke Latin but could not read Greek. His translation was intended to be read aloud to illiterate women also. For the most part, his vocabulary is basic Latin, and he explains any necessary technical terms in simple language. It is unlikely that he would use an obscure regionalism in a translation intended for use throughout the Latin world. Therefore, I will attempt to explain _ubuppa_ as a native Latin word.

J. Pokorny (IEW 1103) recognized a root *ub- 'drängen, (nieder)drücken'(?) forming verbs in Indo-Iranian and Baltic. This looks like the zero-grade of an otherwise unattested PIE *web- 'to press, squeeze' vel sim. In connection with a feeding bottle, *ub- could be the prevocalic combining form of a thematic noun *ubó- 'pressure, squeezing'. If so, _ubuppa_ presumably referred to a squeeze-bottle, as opposed to _titina_, a rigid bottle. If the two terms were entirely synonymous, we would expect M. to use _vel_ and not _aut_, indicating an immaterial choice of words. Naturally the squeezable part of an ubuppa could not be made of glass.

In my post "Kluge's Law in Italic?" (message #68402, 23 Jan 2012; laryngeal notation corrected #68416, 25 Jan), I suggested that such Latin words as _lippus_, _siccus_, _glittus_, and _mittere_ owe their geminated tenues to the operation of Kluge's assimilation. So far I have found no good reason to think otherwise. Thus, if the PIE root *h2/3webH- (IEW 1114-5 *webH- 'weben, flechten, knüpfen') formed a noun *h2/3ubH-néh2 'woven vessel', this would become Proto-Italic *uppa:, and this term could have distinguished a flexible vessel from a rigid one. Then *ubuppa: 'squeezable vessel' could have become specialized as 'squeezable feeding bottle'. The object itself would seldom have found its way out of the realm of nursing mothers and the obstetrices who attended to them, and so the word could easily have been overlooked by the male Latin grammarians of the classical period.


*****

 

An alternative explanation of the first element which does not depend on the obscure zero-grade PIE *ub- has suggested itself.  Rather than native Latin, _ubuppa_ may be of P-Italic origin, like such words as _bo:s_, _lupus_, _scro:fa_, and _forfex_, which have non-Q-Italic consonantism.  If so, the first element could be the P-Italic reflex of *ugW(o)-, from the zero-grade of PIE *wegW- 'wet' (IEW 1118).  The Proto-Italic compound *ugWuppa: would then signify 'woven vessel for liquids', 'wet bag', or the like.  Technologically, this is preferable to 'squeezable bottle', since it is unlikely that such things could have been produced before rubber was known.  But a bag woven around a dead animal's removed stomach, or a woven skin bag, would have been technically feasible without rubber.

 

Beside *ugW-, Pokorny allowed a long zero-grade *u:gW- in order to accommodate Latin _u:vidus_ and the like.  This probably acquired u:- from its more common synonym _u:midus_, not a variant zero-grade.  That _u:midus_ represents earlier *ugW-sm-ido- was accepted by Ernout-Meillet, Walde-Hofmann, and Watkins, but doubted by Weiss (cited by de Vaan, EDL s.v. _u:meo:_), who proposed a different root.  This controversy about _u:midus_ (and indirectly about _u:vidus_) has no bearing on _ubuppa_, however.  It, or its protoform *ugWuppa:, would have been formed long before there was any analogical pressure to lengthen the root-vowel in _u:vidus_.

*****

 

In hindsight, neither a bag woven around a stomach nor a bag woven of skins is likely to have given its name to _ubuppa_, which was after all rigid, a _vasculum vitreum_.  A better explanation is that bags woven from flexible twigs or whatever served as frameworks for earthen vessels, which upon firing became suitable for holding liquids.  A Proto-Italic *uppa: 'woven container' thus became an *ugWuppa: '(woven) vessel for liquids', P-Italic *ubuppa:, borrowed into Latin as _ubuppa_.

 

Vannetais Breton _offen_ f. 'stone trough' reflects Proto-Celtic *uppa: (J. Loth, Rev. Celt. 43:410, cited by Walde-Hofmann, LEW s.v. _aulla_).  In my view this is identical to the Proto-Italic word which I postulated, since I regard Kluge's Law (= Stokes' Law in Celtic) as operating in Old Western IE, the common ancestor of Germanic, Celtic, Italic, Venetic, and Ligurian.  The Celtic sense-development of *uppa: was presumably 'woven container' > '(clay-covered woven) container for liquids' > 'trough'.

 

Latin _faluppa_ 'piece of thread, bit of straw, little chip', although not attested before 10th-century glosses, may be formed like _ubuppa_.  Among the Romance reflexes (W. Meyer-Lübke, REW 3173.2) is Brescian _falopa_ 'defect in weaving'.  If this continues the original sense, the word could have come to signify 'missing piece of thread', whence 'detached piece of thread' and generally 'small thing of minimal value' as attested in medieval Latin.  De Vaan groups Latin _fallo:_ 'I deceive, mislead' with Greek _sphállo:_ 'I trip up, cause to stumble' (intr. aor. _esphále:n_ 'I stumbled') and Sanskrit _skhalate_ 'stumbles, fails' as reflexes of PIE *(s)gWHh2(e)l-(n)-.  Greek and Latin evidently have a transitive nasal present.  Greek also has deverbal nouns _sphálma_, _sphalmós_, _sphálsis_ 'trip, stumble, misstep, mistake, fault'.  A Proto-Italic deverbal *fal(o)- 'mis-' prefixed to *uppa: would have yielded *faluppa: 'misweave, (instance of) faulty weaving' which is practically the same sense as Bresc. _falopa_.

 

Meyer-Lübke took Old French _enveloper_, Italian _inviluppare_ 'to wrap up', and related words, as resulting from contamination of _faluppa_ with Lat. _volvo:_ 'I roll' and its derivatives (REW 3173.3).  The vocalism however fits *velupp- better than *volupp-, and no contamination is necessary if Latin borrowed *veluppa from P-Italic (since a native word would require *voluppa with /l/ pinguis; that -elu- was enunciable is shown by the compound conjunction _velut_ 'as though').  I posit a Proto-Italic *wel(o)- 'rolling, ball-shaped, round' prefixed in *weluppa: 'round woven object, cocoon, chrysalis'.  In this scenario, Vulgar Latin had *inveluppa:re (*invil-) 'to enclose by weaving, enclose as with a cocoon, envelop' and *de:veluppa:re, *exveluppa:re 'to emerge as from a cocoon, develop, sviluppare'.  Italian _viluppo_ 'tangle, ravel', which M.-L. places first as the most primitive form, is in my view a back-formation from one of the verbs.

 

Although no phonetic evidence suggests itself, I suspect that _faluppa_, like _ubuppa_ and *veluppa, was a P-Italic loanword to Latin.  The classical language had very few compounds like _angiportum_ 'narrow street', already an archaism.  Native speakers preferred separate words to compounds.

 

 
DGK