Re: egg

 

 

Michiel de Vaan derives Latin _o:vum_ 'egg' through Proto-Italic *o:w(j)om from PIE *h2o:uio- 'belonging to a bird', a vr.ddhi-derivative of *h2eui- (surface-true *h2awi-) 'bird' (EDL s.vv. _o:vum_, _avis_).  De Vaan gives no basis for the parenthetization of *(j) in Italic which presumably led to its absence from Latin.  He suggests that an earlier full ablaut-alternation /e/:/o/:zero has disappeared from the paradigm of 'bird', with the /o:/-grade of 'egg' lengthened from the vanished /o/-grade stem *h2oui- of 'bird'.

 

R.S.P. Beekes invokes such an /o/-grade to explain Greek _oio:nós_ 'large bird of prey; augurial bird; omen' as a thematization of *h2oui-o:n (Die Sprache 18:21, 1972; EDG s.v.).  A tetrasyllabic _oïo:nós_ occurs in Trypho and possibly Alcman, lending support to an earlier *owio:n-.  Although Beekes agrees with de Vaan in reconstructing *h2o:uio- for 'egg' (EDG s.v. _o:ión_), he does not endorse connecting it with 'bird' through vr.ddhi.  Instead he refers the reader to "the extensive and careful treatment" by Jochem Schindler ("Die idg. Wörter für 'Vogel' und 'Ei'", Die Sprache 15:144-67, 1969).  De Vaan, however, explicitly rejects Schindler's reconstruction *o:-h2uiom (with preposition *o:) 'which is near the bird'.

 

Among the cognates of _avis_ cited by de Vaan, the Celtic forms are dubious and do not identify the presumed laryngeal, and the Indo-Iranian forms begin with v-, evidently from zero-grade with laryngeal apocope.  Armenian _haw_ 'bird, chicken' seems to point to *h2aw-, but as Julius Pokorny noted (IEW 86), it could also continue the zero-grade of his *po:u- 'young of an animal', from which he derived Balto-Slavic *puta: 'bird' (IEW 842-3), and other protoforms are also compatible.  That leaves only Greek _aietós_ (Attic _a:etós_, Pergaian _aibetós_) 'eagle' as corroboration for /h2/-anlaut.

 

Beekes takes the Common Greek *aiwetós 'eagle' as a metathesis from *awjeto-, and compares _niphetós_ 'shower of snow, snowstorm', _puretós_ 'fiery heat, fever' for the suffix.  Here he follows Pierre Chantraine (DELG s.v. _aietós_), who in turn followed Eduard Schwyzer (Gr. Gr. 1:501) and Wilhelm Schulze (Kl. Schr. 75).  Despite the impressive tradition, this is very difficult to understand semantically.  The Greeks regarded the eagle as the greatest of the birds of prey.  Homer described it as _Kronío:ni ... phíltatos oio:nôn_ 'dearest of the birds of prey to Zeus' (Il. 24:293), Aeschylus as _oio:nôn basileús_ 'king of the birds of prey' (Agam. 114), and Pindar as _arkhòs oio:nôn_ 'leader of the birds of prey' (Pyth. 1:14).  That the Greeks should have called the eagle a mere 'episode of birdliness' (or whatever 'bird' plus the denominal suffix -etós is supposed to mean here) defies good sense.

 

I prefer to take _ai(w)etós_ as a deverbative like _herpetón_ 'creeping animal, reptile, snake' from _hérpo:_ 'I creep', _daketón_ 'biting animal' from _dákno:_ 'I bite' (2nd aor. _édakon_).  The heroic name _Aía:s_ 'Ajax', gen. _Aíantos_ (nom. _Aiwas_ on Corinthian and Etruscan vases) is evidently another deverbative from the same obsolete verb; the formation is parallel to _a-káma:s_ 'tireless, indefatigable' from _kámno:_ 'I get tired'.  The root of _ai(w)etós_ and _Aí(w)a:s_ is probably *h2/4aiw- 'to be full of vital force, lively, vigorous', identical to Pokorny's nominal root *aiw-, *aju- 'Lebenskraft' (IEW 17-8).  This is not directly attested as a verb, but appears in Greek as _aió:n_ '(life)time' and _aién_ 'ever'.  The original sense of both _ai(w)etós_ and _Aí(w)a:s_ is thus 'lively, vigorous' vel sim., the former more suitable for animals, the latter for humans, and there is no basis for connecting the former with Latin _avis_.

 

Thus the 'egg' and 'bird' words under study need not have begun with *h2.  A root *h3eh1-, surface-true *h3oh1-, can serve instead.  Greek _o:ión_ in this view continues the full-grade derivative *h3oh1-jóm 'egg', which never contained *w and also underlies Iranian *a:ja- (Persian _xa:ya_), Slavic *a:je- (Russian _jajcó_), and Germanic *ajja- (discussed below).  'Bird' itself is then an animate compound root-noun *h3h1-wéi-, *h3h1-wí-.  In Italic and Armenian, the *h1 between two consonants was vocalized to *a, then the *h3- was lost in Italic but preserved as h- in Armenian.  In Indo-Iranian, the two preconsonantal laryngeals were lost.  Thus we have Latin _avis_, Umbrian acc. pl. _avif_, abl. pl. _avis_ (P-Italic *awins, *awifs), Armenian _haw_, Sanskrit _vi-_, nom. pl. _vayas_ 'bird'.

 

If in Greek the *h1 between consonants was first vocalized to *e, then colored to *o by the foregoing *h3- which was subsequently lost, *h3h1-wí- would have given *owí- 'bird', differing only in accent from 'sheep' and thus highly susceptible to replacement.  Infixed -o:n- produces 'grandson' from 'son':  _huio:neîs . huiôn huiées_ 'grandsons = sons of sons' (Hesychius).  Homer and others have a thematized _huio:nós_ 'grandson' instead.  Hence *oweîes (i.e. *owéjes = Skt. _vayas_) 'birds' could have produced *owio:neîes, later *o(w)io:neîs 'birds of birds' i.e. 'birds which hunt other birds, birds of prey', with a new thematic singular *o(w)io:nós replacing the old form as in 'grandson'.  This explains _oio:nós_ on the basis of PIE *h3oh1- rather than *h2aw-.

 

If the second element of the 'bird' compound is *wei- 'bend, wind, plait, weave', the literal sense of 'bird' is likely 'nest-weaver', or perhaps more accurately 'home-weaver', distinguishing birds from other fauna which do not establish homes by plaiting or weaving together twigs and debris.  Then the original sense of *h3oh1- was likely 'settle down, establish a home' vel sim., later specialized to birds as 'establish a nest'.  The new thematic noun *h3oh1ó- (replacing the old root-noun) would thus mean '(bird's) nest', and its derivative *h3oh1-jó- 'that which is located in the nest, (bird's) egg'.  The formation is parallel to Germanic *badjaN 'bed' (Old English _bedd_, etc.), representing PIE *bHodH-jóm 'that which is located in the dugout' from *bHedH- 'to dig, excavate'.

 

Latin _o:men_ 'divine sign, prophesy, omen', otherwise an etymological orphan, may fit here as well.  It can continue PIE *h3óh1-mn., originally 'act of establishing a home', later 'act of building a nest'.  Seers who followed the actions of augurial birds would likely restrict the term to the nests of eagles and other such birds.  As all actions of such birds were, at least in principle, prophetic, the building of a nest in a particular place by a mating pair of augurial birds would be regarded as a divine sign.  The term could then have been generalized to any divine sign.

 

Beside the oxytone _o:ión_, Greek has another neuter 'egg' of distinct formation, found in Theognostus as an acc. sg. _ó:eon_ in addition to _o:ión_.  Several other authors use _ó:eon_, and Hesychius glosses an Argive nom./acc. pl. _ó:bea_.  This last form requires *w, unlike the oxytone neuter.  I suspect it was earlier a sigmatic neuter, with the attested thematic neuter back-formed from the plural *ó:we(s)a.  This in turn can be understood, like 'bird', as a compound root-noun, inanimate this time, whose first element, thematic this time, is 'nest'.  The second element is PIE *h2wes- 'remain in place, tarry, stay'.  Thus *h3oh1ó-h2wes- meant 'that which stays in the nest' i.e. 'egg' as opposed to 'bird', since birds leave the nest from time to time.  Loss and absorption of laryngeals would produce *ó::wes-, with the trimoric vowel regularly reduced to an ordinary dimoric long.  This yielded Proto-Greek *ó:wes-, of which I presume only the nom./acc. nt. pl. *ó:wesa was in good use, falling together formally with original sigmatic neuters (i.e. /es/os/-stems) and suppleted by oblique cases of _o:ión_.  After the loss of intervocalic -s-, some dialects back-formed _ó:(w)eon_ from the pl. _ó:(w)ea_ and created a full paradigm which could be used beside _o:ión_, as Theognostus did.

 

I suspect that Proto-Italic similarly had a nom./acc. pl. *o:wesa(:) beside *o:jom.  Very likely the latter could function as a mass-noun 'egg, the edible substance inside the shells', while the former was an unambiguous count-noun used in enumerating eggs.  But such distinctions are not always easily maintained, or necessarily used consistently by all speakers in a given community.  I presume that the Proto-Italic count-noun was reinterpreted as a sigmatic neuter, leading to the nom./acc. sg. *o:wos beside *o:jom.  In Old Latin, this nom./acc. nt. sg. *o:vos would have lost /w/ between like vowels, leading to *o:ös.  This in turn would have been subject to the vowel-correption law, "vocalis ante vocalem corripitur", and shortened to *oös.  When this became Lat. *oüs, it would have acquired secondary /w/, thus *ovus.  A form _ovus_ is indeed attested in the sixth-century Oribasius Latinus, but usually regarded as a masculine by-form of _o:vum_ (thus Alexander Souter, Glossary of Later Latin, indicates _o:vus_).  Nevertheless this short-vowel neuter *ovus can explain why Romance forms such as Italian _uovo_ and Spanish _huevo_ reflect short /o/, not the /o:/ of _o:vum_.  The declension is not a problem, since the Lat. nt. _tempus_ becomes It. _tempo_, Sp. _tiempo_, etc.

 

Plural and oblique singular cases of *ovus would have retained *o:- until they were remodelled after the nom./acc. sg.  I assume that some Latin dialects thus had gen. sg. *o:veris, nom./acc. pl. *o:vera, etc. beside *ovus and the inherited *o:jum, which likely functioned as a mass-noun most of the time.  Contamination very probably produced _o:vum_, which quickly replaced *o:jum in most dialects.  Its force as a collective mass-noun is still apparent in the proverbial expression "ab ovo ad mala", 'from the egg(s) to the apples', i.e. 'from beginning to end' (of a Roman banquet).

 

Literate Romans likely looked down their noses at *ovus as a vulgarism, and used _o:vum_ as a count-noun also (pl. _o:va_).  Yet if my analysis is correct, it was the plebeian *ovus which alone survived into the Romance languages.

 

Germanic words for 'egg' exhibit Verschärfung, which opens up a can of worms involving laryngeal-semivowel clusters and the effect of the inherited PIE accent on their reflexes in Proto-Germanic.  In this post I will not dump out the can.  I intend to examine this matter in detail later, largely following W.P. Lehmann's discussion (PIE Phonology chh. 4-5).  I disagree with Lehmann on two important points.  I reject the formalism of reduced grade, and I accept the Vernerian accent-dependence of the Verschärfung and the related production of *k and *g from certain clusters of laryngeals with *w.  Moreover, I find it practical to identify most of the laryngeals in the words under study, rather than leaving an undetermined laryngeal X in the protoforms.

 

Verschärfung is conventionally represented as the lengthening of intervocalic *-j- and *-w- to *-jj- and *-ww- in certain PGmc lexemes.  In West Germanic, the first semivowel forms a diphthong with the preceding vowel which spares the second, thus *-ij-, *-uw-.  But *-jj- becomes *-ggj- in North Gmc. and *-ddj- in East Gmc., while *-ww- becomes *-ggw- in both NGmc and EGmc.

 

Lehmann proposed a PIE sigmatic neuter *oXy-es- 'egg' (PIE Phon. p. 44).  Outside Germanic this was contracted to *o:y-es- in his view, but PGmc preserved the uncontracted form and developed *aXj-es- > *ajj-es- > *ejj-is- (better *ejj-iz-).  Such a sigmatic neuter is not found outside Gmc., however, and in fact only WGmc plural forms require it:  Old English _æ:gru_ (double-marked dialectal Eng. _eyren_), Old High German _eigir_, Old Saxon gen. pl. _eiiero_.  The sound written as *-jj- was sufficient by itself to produce /j/-umlaut in NGmc and WGmc, and the Gmc. singular forms can be referred to a thematic neuter *ajja-:  Crimean Gothic _ada_ (= Ulfilan Go. *addja), Old Norse _egg_ (borrowed into Middle English), OE _æ:g_, OHG _ei_ (dat. sg. _eiie_).

 

I suspect that PGmc, like Greek and Latin, inherited both *h3oh1-jóm, which was at least a facultative mass-noun, and *h3oh1ó-h2wes-, a strict count-noun.  In my view PIE *-h1j- became PGmc *-jj-, and PIE *-h2w- became PGmc *-k-, only under the same conditions which voiced a fricative by Verner's Law.  Thus we have Gothic _daddjan_, Old Swedish _dæggja_ 'to suckle' against Westfalian _däierrn_ 'to nourish a calf with milk', OHG _ta:an_ 'to suckle' (1sg. _ta:ju_).  The former forms are derived from (or analogical after) PIE *dHoh1-j´- > PGmc *dajj-, the latter from PIE *dHéh1-jV- > PGmc *dæ:jV-.  Likewise, ON _naust_ 'boathouse' and _nóa-tún_ 'town of boats' continue preconsonantal and prevocalic developments of PIE *náh2w-, while the WGmc weak noun OE _naca_, OS _naco_, OHG _nacho_ 'boat', originally an amphikinetic /n/-stem, shows generalization of /k/ from the middle stem, most likely the loc. sg. 'in the boat', (formally) PIE *nah2wéni > PGmc *nakeni, which took over the dat. sg. function in WGmc.

 

In my view then PIE *h3oh1-jóm regularly became PGmc *ajjaN, which is reflected as the singular 'egg' in all three Gmc. branches.  But PIE *h3oh1ó-h2wes- would have become a sigmatic neuter, PGmc *o:wiz-, which seems to have gone out of use.  Nevertheless its declension apparently contaminated the plural of 'egg', which for a time must have exhibited sg. *ajja- beside pl. *ajjiz-, the latter surviving into WGmc.

 

Douglas G. Kilday