Main

The prehistory of the Baltic preterit system is quite unclear despite its synchronic simplicity. The Baltic languages (Lithuanian, Latvian, and Old Prussian) have only two formations for the simple preterit:

• ā-preterit

e.g., Lith. 3sg./pl. pir̃ko, Latv. 1sg. pìrku ‘bought’, Old Prussian 3sg./pl. kūra 'built'; and

• ē-preterit

e.g., Lith. 3sg./pl. vẽdė, Latv. 1sg. vedu(ē) 'led', OPru. 3sg. weddē[din] ‘brought [them]’.

The Lithuanian dictionary forms: infinitive, 3sg./pl. present form, 3sg./pl. preterit form,
(singular and plural forms are the same in the 3rd person in Baltic)

Latvian dictionary forms: infinitive, 1sg. present form, 1sg. preterit form.

For old Prussian, often these dictionary forms are not all attested, and only the attested forms are given.

The synchronic distribution of the ā- and ē-preterit is to a great degree predictable based on their root structure & present stem formation as shown by Schmid (1966). For example,

• when the root has the CeC-structure (C = consonant) and the present stem is CeC-a-, it takes an ē-preterit (CeC-ē),

examples: inf. dègti, 3sg./pl. pres. dẽga, 3sg./pl. pret. dẽgė 'to burn (tr.)', vèsti, vẽda, vẽdė 'to lead', nèšti, nẽša, nẽšė 'to carry';

• when the root has the CeR-structure (R = resonant) and the present stem is CeR-a-, it takes an ē-preterit (CeR-ē),

examples: gim̃ti, gẽma, gìmė 'to be born', giñti, gẽna, gìnė 'to chase';

• when the root has the CeRC-structure and the present stem is CeRC-a-, it takes an ā-preterit (CRC-ā-),

examples: kir̃pti, ker̃pa, kir̃po 'to trim', mìlžti, mèlža, mìlžo 'to milk', lìkti, liẽka, lìko 'to remain',

etc., etc.....

But it is unclear

how the distribution of ā-preterit and ē-preterit has developed,

and how the ē-preterit in particular developed.

In addition, the ē-preterit includes a problematic category called “long-vowel preterit” (e.g., Lith. emė ‘took’).

The aim of this proposed project is to clarify the prehistory of the Baltic preterit system:

• the mechanism of the distribution of ā-preterit and ē-preterit,

• and the historical origins of the ā- and ē-preterit.

Previous studies have approached the Baltic preterit system from various perspectives. Through the previous studies, it appears to me to be two main aspects in the diachronic development of the Baltic preterit (or the whole verbal) system:

• Diathesis-based development
• Aktionsart-based development.

Diathesis-based development:

Some Baltic verbs show a distribution of intransitive ā-preterit and transitive ē-preterit (Endzelīns 1923: 567ff., 667; Stang 1966: 377; and Kølln 1969). The intransitive and transitive pairs that show this distribution include:

• intransitive nasal-present verbs (with the full-grade root!) and transitive simple thematic verbs:

intransitive	transitive
OLith. 1sg. pret. ussidegau ‘to get inflammation’	Lith. dègti, dẽga, dẽgė 'to burn' (tr./intr.)
Latv. dial. 1sg. pret. dagu / degu 'to burn'	Latv. dial. 1sg. pret. dedzu / d’edźu(ē) ‘to burn’
Lith. kèpti, kem̃pa, kẽpo 'to become dry'	Lith. kèpti, kẽpa, kẽpė 'to bake'



• pairs of intransitive nasal-infix present verbs transitive ia-present verbs:

intransitive	transitive
Lith.bìrti, bỹra / bìrsta, bìro 'to crumble'	Lith. berti, bẽria, bėrė 'to scatter'
Lith. mùkti, muñka, mùko ‘to fall off’	Lith. maũkti, maũkia, maũkė 'to release'
Lith. trùkti, truñka, trùko 'to continue'	Lith. tráukti, tráukia, tráukė 'to drag'.

The following things have been pointed out:

• Some ā-preterits have middle-intransitive meanings, and they are comparable to Slavic thematic aorist (Stang 1942: 63-64).

• Such ā-preterit and Slavic thematic aorist may have developed from PIE middle root-aorist (Villanueva Svensson 2011b).

• Their paired nasal-infix presents, which are intransitive in Germanic, Slavic and Baltic, in contrast to other IE languages, may have developed from the PIE middle paradigm of nasal-infix presents (id.).

Aktionsart-based development:

It has been pointed out that the "aspects" of verbs played an important role in the development of the Balto-Slavic verbs. "Aspect" refers to the phases of an action or event denoted by a given verb:

A verb can take a form that denotes the start of an action or event, e.g., "start to rain, start to think....", or the ongoing action or event like "be raining, in the middle of thinking...." or just the whole action without any particular specification, e.g., "rain, think".

terminology		approximate meaning
inchoative		start to do / happen,
imperfective / durative		be doing / be going on
aorist		did / happened
perfect		have done / happened.

But it is also known that the lexical meanings of verbal roots have an aspectual feature as a part of their semantic features in PIE. This sort of aspect-like feature that a given verbal root already has as a lexical information is called Aktionsart.

For example, the PIE verbal root for "to be" was *h₁es-. This root formed a root present, i.e., 1sg. *h₁es-mi, 2sg. h₁es-si, 3sg. *h₁es-ti, etc., the past tense form in the imperfect aspect 1sg. (h₁e)-h₁es-m, 3sg. *(h₁e-)h₁es-t, etc., but does not seem to have formed a root aorist.

This can be explained like this: the lexical meaning of this verb "to be" implies a durative feature, i.e., *h₁es- was a durative / present root. This feature was compatible with the root-present, imperfect, but not with the root-aorist formation, which typically conveyed the completion of an action / event as a whole, as though it happened "punctually".

Any action / event in reality would take a certain duration of time to take place. But this is not what the aorist formation expresses. Instead, it expresses the whole action / event as if it had occupied only a moment on the time scale.

In this way, the Aktionsart of a verbal root and the grammatical aspect forms that the root can form have correlations.

In Barton's (1980: 250) and Aitzetmüller's (1991: 166) works, descriptions of how the Balto-Slavic verbal system developed, being motivated by this Aktionsart is found.

In PIE, the verbs with aoristic Aktionsart formed root aorist with no specific markers. But their present stems were characterized by some specific formations (like "thematic" present, reduplicated present, etc.) that would mark the "durative" meaning of the present. Below are the examples from an aoristic root *der- 'to tear' (cf. LIV2 119). The listed forms are in the 3rd person singular.

	Punctual		Durative
Pres.	---		them.-pres. *der-e-ti
Aor.	root-aor. *der-t		---

In Proto-Balto-Slavic, the characterized "punctual" (→ Sl. "perfective") presents and "durative" (→ Sl. "imperfective") aorists were created.

	Punctual			Durative
pres.	*dire-ti		them.-pres.	*dere-ti
root-aor.	*der-t		ā-aor.	*dirā-t

These new forms are characterized with the zero-grade. The "durative" aorist is typically marked with the -ā-suffix, and called the Balto-Slavic ā-aorist.

The OCS verbs je̢ti, imo̢ 'to take' and imati, jemljo̢ 'to have' may well represent this development (cf. Aitzetmüller 1962).

	Perfective			Imperfective
pres.	*ime-ti		them.-pres.	*em(y)e-ti
root-aor.	*em-t		ā-aor.	†imā-t (not attested)

The Balto-Slavic ā-aorist is the origin of some Baltic ā-preterit (Stang 1942: 75ff, Kølln 1969, Villanueva Svensson 2020). But their developments differ depending on their root-structure (→ see below).

There are further more aspects in the development of the Baltic preterit system. The verbs have been influenced by phonological and morphological patterns and semantic features in their development, and these factors seem to affect each other. Therefore, the data are collected and assorted in accordance with some phonological, morphological and semantic groups as below. Relevant previous studies are also mentioned below.

• transitive ē-preterit
  → Endzelins (1923), Stang (1942, 1966), Barton (1980), Vilanueva Svensson (2005)
  • transitive ē-preterit to ia-present
    • long-vowel preterit
      → Kortlandt (1985), Jasanoff (2012),

      → Yamazaki(2019a)
      • with tone variation
      → Larsson (2011), Villanueva Svensson (2014)
      • without tone variation
    • no lengthening in the root
  • ē-preterit to simple thematic present (without tone-variation)
  → Konstatinova & Kapsevičienė (1989)
  about the verb Lith. duoti, duoda, davė 'to give'
  → Yamazaki (2019b)
• middle-intransitive ē-preterit (to simple thematic present or sta-present)
→ Barton (1980), Villanueva Svensson (2005);
• intransitive ā-preterit
→ Stang (1942: 61-62)
• transitive ā-preterit
→ Yamazaki (2021/2023)
• beside, causative-iteratives in -yti- always take the ē-preterit

For the whole preterit system including transitive and intransitive ā-preterit, a manuscript* is underway. Bibliography list is found here.

My investigation so far indicates (cf. presentation slides presented at The Congress of Balticists 2021):

• Baltic transitivity contrast in the ā- and ē-preterits corresponds to diathetic contrast in other IE languages (Indo-Iranian and Tocharian):
  

  intransitive ā-preterit	:	IE middle
  transitive ē-preterit	:	IE active

• On the other hand, the Aktionsart-based development also left relics in some transitive ā-preterits.
  


• Root-structure seems to have played a role in the Aktionsart-based development:
  • CeRC-root → preserving the ablaut pattern of the durative present & aorist:
  
  

  	Baltic		Slavic
  *durat.	Lith. kirpti, kerpa, kirpo		---
  	'to trim'
  *punc.	---		OCS po-črěti, -crъpo̢
  			'to scoop'

  
  
  • CeR-root and transitive → they joined the transitive ia-present verbs (paired to the ē-preterit, see above):
  
  

  	Baltic		Slavic
  *durat.	dirti, dera/diria, dyrė		OCS dirati, dero̢, diraxъ
  	'to flay'		'to tear, flay'
  *punc.	---		CS drěti, dьro̢ 'to tear'

  
  
  • The Balto-Slavic ā-aorist with CeR-roots possibly joined the Baltic ē-preterit. In such a case, the ā-stem was introduced due to an analogical factor, as described in Barton (1980), Villanueva Svensson (2005).
  
  
• The middle-intransitive thematic aorist and the durative / iterative ā-aorist merged to the ā-preterit in Baltic. A pivotal form may have caused the merger.
  

  PIE	3sg. mid. root-aor.	3pl. mid. root-aor.
  	*muk-(e)t ‘fell off’	*muk-onto	---	---
  	⇓	⇓
  PBS	3sg. them-aor.	3pl. them-aor.	3sg. ā-aor.	3pl. ā-aor.
  	*muk-e-(t)	*muk-o-nt	*kirp-ā-(t)	*kirp-ā-nt
  			‘was clipping’
  	⇓	⇓	⇓	⇓
  pre-PB	*muka	*muk-an	*kirp-ā	*kirp-an

  
  The 3pl. forms might have provided pivot forms:
  
  3sg. *kirp-ā : 3pl. *kirp-an = x : *muk-an, where x = *mukā.

Research papers:

• The presence / absence of the tone variation among the ē-preterit may depend on their origins: → Yamazaki(2019a)
  • previous root-aorist, or barytone preterit stems → no variation;
  • mobile preterit stems incl. imprefect of Narten-presents → tone variation


• Middle-intransitive ē-preterit: intransitive verbs usually take the ā-preterit, which is paired with the nasal-infix present or sta-present. Middle-intransitive verbs are exceptional, in that they take the ē-preterit despite their intransitive meaning. They must have developed in East Baltic as a peculiar category.
  • Yamazaki 2021 discusses two Baltic middle-intransitive verbs, Lith. mir̃ti / Latv. mirt `to die' and Lith. mir̃šti / Latv. mìrst `to forget', and the historical origins of their preterit stems.
  
  
• The ā-preterit developed from at least two major categories: the Proto-Balto-Slavic thematic aorist (≺ IE *middle root-aorist, cf. Stang 1942, Villanueva Svensson 2011b); and the Balto-Slavic "durative" ā-aorist (cf. Barton 1980: 250 and Aitzetmüller 1991: 166).
  Recently, Villanueva Svensson (2020) proposed that the BSl ā-aorist may have developed from the iteratives in *-eh₂-.
  This may be one of the reasons that the ā-preterit includes both intransitive and transitive verbs. Another reason may be that some transitive ā-preterit developed from the previous middle paradigm. When the (oppositional) middle inflection was shifted to the reflexive construction, their middle function was reinforced by the reflexive construction, while the forms without the reflexive marker obtained transitive meanings. This proposal is found in Yamazaki (2021/2023).


• anomalous preterit forms have interesting individual prehistories:
  • Lith. dãvė to dúoti 'to give' → Yamazaki (2019b)
  
  
• The development of the pattern of the ē-preterits paired with the ia-presents can be viewed from a perspective a little different from that of the sound law and analogy. A paper addressing this topic is to appear in an edited volume of the Fachtagung of Indo-European Society that was supposed to take place in Zurich 2020.


• I was working for the edited volume of Fachtagung Zurich 2020 as a contributor and as an editor / typesetter. I am grateful to Prof. Paul Widmer, the host at the Department of Comparative Language Science, for his generous support and providing me with this opportunity to collaborate with his team.
  • Update (2022-12-15)
    The volume was published in the fall 2022:
    Sommer, Florian, Karin Stüber, Paul Widmer, and Yoko Yamazaki (eds.) Indogermanische Morphologie in erweiterter Sicht. Innsbruck: Institut für Sprachwissenschaft der Universität 2022.
  • Yoko Yamazaki (2022) "The Baltic transitive ia-presents and their paired preterits in the ē- or *iyā-stem — Between a conditioned allomorph and an independent morpheme" In: Sommer et al. (eds.) Indogermanische Morphologie in erweiterter Sicht, pp. 413–430.

Bibliography

List of bibliography.

Lectures:

• "Diathetic problem of the Baltic ā-preterits to the simple thematic presents" October 14th, 2021, XIII Starptautiskā Baltistu Kongress, University of Latvia. [Presentation slides].


• "Investigation on the prehistory of the Baltic preterit system" November 28th 2019, Språkhistoriska seminariet, Stockholm University. Available on youtube.