Search

Stang’s law is an ambiguous concept. On the one hand, it refers to a retraction of the stress from non-acute long vowels to the preceding syllable in Slavic. On the other hand, it refers to the development of acc.sg. Vedic dyā́m ‘sky’, gā́m ‘bull, cow’, Greek Ζῆν(α), βῶν, from *dieum, *gwoum. Stang derived the long vowel from the diphthong before the tautosyllabic nasal consonant. Alternatively, the long vowel can be attributed to monosyllabic lengthening followed by the loss of *‑u‑ before the tautosyllabic nasal.I have proposed that *‑u‑ was lost before word-final *‑m at a stage before the monosyllabic lengthening, yielding *diēm < *diem < *dieum and *gwēH3m < *gweH3m < *gweH3um. Latvian gùovs reflects the acc.sg. form *gwēH3m with loss of the laryngeal after the long vowel. The paradigm with a full grade suffix *‑eH2‑ and the loss of laryngeals before final *‑m were dialectal Indo-European innovations. Nasal vowels in final syllables lost their nasality in East Baltic.The loss of *‑u‑ before *‑m in Vedic gā́m and Greek βῶν, the rise of the long vowel in these forms, the loss of the laryngeal in Latvian gùovs, the generalization of the full grade *‑eH2‑ in the paradigm, the loss of laryngeals before *‑m, and the loss of nasality in East Baltic final syllables are all chronologically distinct developments, none of which can appropriately be called Stang’s law. It is therefore preferable to use this term only in reference to the retraction of the stress from non-acute long vowels in Slavic, which is the basis of modern Slavic accentology.

The basis of linguistic reconstruction is the comparative method, which starts from the assumption that there is “a stronger affinity, both in the roots of verbs and in the forms of grammar, than could possibly have been produced by accident”, implying the existence of a common source (thus Sir William Jones in 1786). It follows that there must be a possible sequence of developments from the reconstructed system to the attested data. These developments must have been either phonetically regular or analogical. The latter type of change requires a model and a motivation. A theory which does not account for the data in terms of sound laws and well-motivated analogical changes is not a linguistic reconstruction but philosophical speculation.The pre-laryngealist idea that any Proto-Indo-European long vowel became acute in Balto-Slavic is a typical example of philosophical speculation contradicted by the comparative evidence. Other examples are spontaneous glottalization (Jasanoff’s “acute assignment”, unattested anywhere in the world), Jasanoff’s trimoraic long vowels, Eichner’s law, Osthoff’s law, and Szemerényi’s law, which is an instance of circular reasoning. The Balto-Slavic acute continues the Proto-Indo-European laryngeals and the glottalic feature of the traditional Proto-Indo-European “unaspirated voiced” obstruents (Winter’s law). My reconstruction of Proto-Indo-European glottalic obstruents is based on direct evidence from Indo-Iranian, Armenian, Baltic and Germanic and indirect evidence from Indo-Iranian, Greek, Latin and Slavic.

The closest relatives of Balto-Slavic are Albanian and Indo-Iranian. Together with Armenian and Thracian, these are the satǝm languages, which together with Greek and Phrygian constitute the eastern part of Classic Indo-European.The obvious common development in this area is the satǝm palatalization, which did not affect Greek and Phrygian. Indo-Iranian was separated from the other satǝm languages by the depalatalization of palatovelars before resonants in the latter.Proto-Indo-European had a threefold distinction between fortis, glottalic lenis, and plain lenis obstruents, all of them voiceless, e.g. *t [t:], *d [ť], *dh [t]. In the Classic Indo-European languages (after their separation from Anatolian and Tocharian), the lenis obstruents became voiced [ʔd], [d], while the fortis remained voiceless [t]. This system was best preserved in Indo-Iranian, Balto-Slavic and Albanian.Another development that seems to be dialectal Indo-European is the retraction of *s to *ṣ after *i, *u, *r, *k in Indo-Iranian, Balto-Slavic, Albanian and Armenian. However, this retraction cannot be dated and may have affected any part of the IndoEuropean dialects. This may also explain the Hittite reflex š of PIE *s.The large majority of special correspondences between Balto-Slavic and IndoIranian are archaisms, not innovations. This is important because it implies that a comparison of Balto-Slavic with Indo-Iranian leads to a reconstruction of an early stage of Indo-European.

This paper addresses the relationship of the Lith. nom. pl. endings ‑ai (‑aĩ) and ‑i, ‑ie (‑ì, ‑ìe) to each other and to their common source, the PIE pronominal nom. pl. in *‑oi. It is argued that the Proto-BSl. preform was *‑ai, with acuteness analogically taken from the corresponding nominal ending *‑ōs. Proto-BSl. *‑ai, which developed regularly to ‑i in Slavic and to ‑ai in Old Prussian, had two reflexes in Lithuanian. The phonologically regular treatment is seen in the pronominal and adjectival ending ‑i(e), which developed from *‑ai to *‑ẹ̄ via the presumed intermediate stage *‑ɛi. The nominal ending ‑ai is a morphological treatment of the same sequence, with a) *‑ɛi remade to *‑ai under the influence of paradigmatically related forms in *‑a‑, and b) acuteness subsequently lost in a final syllable. The second step runs counter to standard opinion, which holds that underlying acute diphthongs in final syllables trigger Saussure’s Law and retain their acuteness in Žemaitian, as in the ā‑present forms 1 sg. sakaũ, 2 sg. sakaĩ (Žem. sakâu, sakā̂). It is maintained here, however, that the verbal endings that appear as ‑aũ and ‑aĩ in standard Lithuanian are not historically diphthongs at all, but contracted sequences that arose after the “de-acuting” of the inherited nom. pl. in *‑ai to ‑ai and before Saussure’s Law. A byproduct of this explanation is that other descriptively non-acute diphthongs in final syllables, notably the dat. sg. in ‑ui (*‑ōi) and the instr. pl. in ‑ais (*‑ōis), can also be assumed to have been originally acute, as their etymologies require.

There are two chronological layers of metatonical circumflex in monosyllables, viz. an early Balto-Slavic layer which is reflected e.g. in Lith. dė̃s, jõs, duõs and a recent Aukštaitian layer which is found e.g. in nom.pl. tiẽ, acc.pl. tuõs, inst.sg. tuõ. Leskien’s law was younger than the former but older than the latter. This analysis is not based on a comparison with Slavic or Indo-European but on the internal evidence of the East Baltic languages.The Baltic future represents two Indo-European paradigms, viz. an s‑present with accentual mobility between the suffix and the ending and an s‑aorist with fixed stress on the root and monosyllabic lengthening in the 2nd and 3rd sg. forms. Both of these formations have exact correspondences in the Old Irish subjunctive. They must have existed side by side in Proto-Baltic in view of Prussian teīks ‘make!’ beside postāsei ‘you will become’.The circumflex of Latvian sā̀ls and gùovs shows metatonical length as a result of monosyllabic lengthening. There is no evidence for a PIE phoneme *a in sā̀ls, zùoss and nãss, nor for the vowel *e in the PIE paradigm of gùovs, nor for a PIE paradigm with fixed stress in the case of gùovs, nãss and zvrs, nor for a generalization of the original nom.sg. instead of acc.sg. accentuation in sā̀ls and nãss.

In Baltistica 46(1), Miguel Villanueva Svensson presents a defence of the so-called “traditional” view on the development of long vowels in Balto-Slavic, in opposition to the views of the “Leiden school” (see Frederik Kortlandt’s Long vowels in Balto-Slavic, Baltistica 21(2)). In Baltistica 47(2), Tijmen Pronk replies to Villanueva Svensson’s points.In this paper, the author would like to take some of the items discussed by Villanueva Svensson and Pronk and hold them against his own interpretation of the facts, which differs both from the Leiden and the “traditional” views.The following items are discussed:- Nom sg. of nouns ending in a resonant- Lithuanian -ė- Nouns with long root vowel: root nouns?- Nouns with long root vowel: vowel stems- Nouns with an acute diphthong- Pronominals- The Slavic (sigmatic) aorist- The Baltic ē-preterit- The Baltic future- Narten presents- ieškóti- Intensives /iteratives- je-presents

The traditionally assumed intermediate Balto-Slavonic stage after the break up of Proto-Indo-European can be additionally supported by two highly specific and thus potentially exclusive developments at the end of a word. The first development is the loss of the PIE word-final short *i after a long vowel followed by a labial consonant. This sound change must have occurred later than the secondary shortening of long vowels before word-final PIE *m which is shared by Celtic. The second sound law is the raising of stressed PIE *o to *u in word-final position before consonants both in Baltic and Slavonic. The postulated sound changes are established by means of an in deep analysis of the relevant endings in the inflection of Baltic and Slavonic nouns, pronouns and adjectives.

We arrive at the following relative chronology:1. Depalatalization before syllabic *r̥ (Proto-Indo-European),2. Depalatalization before syllabic *l̥ (Balto-Slavic, Albanian, Armenian),3a. Rise of an epenthetic vowel before syllabic nasals (Balto-Slavic, Armenian),3b. Depalatalization before syllabic nasals (Albanian),4. Rise of an epenthetic vowel before syllabic nasals (Albanian).

The attractiveness of projecting a variety of formations back in time lies in the freedom it allows the investigator to choose between different reconstructions in accordance with his theoretical preconceptions. The history of Indo-European reconstruction can to a large extent be seen as a gradual limitation of this freedom. It is therefore advisable to be cautious when a new theory enlarges rather than restricts the number of possible reconstructions.

The establishment of dominance patterns does not automatically yield insights into the history of accentual paradigms. As in the case of segmental features, it is necessary first to identify the results of analogical developments and to separate them from the original regularities. Both apophony and metatony became highly productive in Latvian and Lithuanian word formation, leading to complex morphonological relationships.