The Indo-European Verb

Proceedings of the Conference of the Society

for Indo-European Studies, .Los Angeles 13-15 September 2010
Edited by
H. Craig Melchert
Wiesbaden 2012
Reichert Verlag
Table of Contents
Foreword vn
BENEDETTI, Marina: Valency Alternations with Perception Verbs m Indo-European
Languages 1-6
BOZZONE, Chiara: The PIE Subjunctive: Function and Development 7-18
DAHL, Eystein: Towards an Account of the Semantics of the PIE Imperative 19-28
DAUES, Alexandra: Hittite Verbs in -ssa-: Can a Function Be Recognized? 29-41
Dr G!OVINE, Paolo: The Function of *a-Ablaut in the PIE Verbal System 43-50
ESKA, Joseph F.: Absolute and Conjunct, Cowgill and Apocope 51-59
GARCiA CASTILLERO, Carlos: The Old Irish Paradigm of Clause Types 61-72
GARCiA RAMON, Jose Luis: Aspect and Mood in Indo-European Reconstruction 73-85
HACKSTEIN, Olav: When Words Coalesce: Chunking and Morphophonemic Extension 87-104
HILL, Eugen, and Michael FROTSCHER: The Accentuation of Old ludic Reduplicated (3rd
Class) Presents 105-114
HOCK, Hans Henrich: Phrasal Prosody and the Indo-European Verb 115-126
JASANOFF, Jay H.: Long-vowel Preterites in Indo-European 127-135
KIM, Ronald I.: Unus testis, unicus testis? The Ablaut of Root Aorists in Tocharian and
Indo-European 13 7-149
KLOEKHORST, Alwin: Hittite "ii/e"-ablauting Verbs 151-160
KOCHAROV, Petr: Perfect Reduplication in Late Indo-European 161-165
KOLLIGAN, Daniel: Patterns of Suppletion in Classical Armenian: The Case of Motion
Verbs 167-177
KRASUKHIN, Konstantin G.: Indo-European Conjugation: History and Pre-History 179-189
KROONEN, Guus: Reflections on the a/zero-Ablaut in the Germanic Iterative Verbs 191-200
KOMMEL, Martin Joachim: The Inflection of the Hittite Verb Class of mema/i- 201-208
LEI-INERT, Christian: Anmerkungen zum homerischen Augment 209-212
LDI-IR, Rosemarie: Ereignistyp und Diathesenwechsel im Indogermanischen 213-224
MAJER, Marek: An Archaic Indo-European Verbal Form in the Slavic Generalizing Particle
*-ildo? 225-234
MALZAHN, Melanie: Archaism and Innovation in the Tocharian Verbal System: The Case
of Valency and the Case for a Conspiracy Theory 235-240
OETTINGER, Norbert: Das Verhaltnis von nominaler und verbaler Reduplikation im
Indogermanischen und Anatolischen 241-246
PEYROT, Michael: e-grade in Tocharian Verbal Morphology 247-256
PINAULT, Georges-Jean: Interpretation of the Tocharian Subjunctive of Class III 257-265
vi
POOTH, Roland A.: Zum Auflwmmen transitiver Verben im frlihen Vedischcn am Beispiel 267-284
lr
RASMUSSEN, .Tens E.: The Origin of the Albanian Mediopassive 285-288
REINHART, Johannes: Inheritance or Innovation in the Proto-Slavic Verb: the Ending -mo
(1st Person Plural) 289-294
SCIIEUNGRABER, Corinna: Nasal Suffix Verbs in Germanic and KLUGE's Law 295-304
Sow A, Wojciech: The Phrygian Middle 305-313
DE V AAN, Michie!: Latin Deverbal Presents in -a- 315-332
VILLANUEVA SVENSSON, Miguel: The Ablaut of the Middle Root Athematic Presents in
Indo-European 333-342
YOSHIDA, Kazuhiko: Notes on Cuneiform Luvian Verbs in *-yelo- 343-351
ZIEGLER, Sabine: Zur Konzeption moderner Worterbiicher: Probleme der Philologie und
der Lexikographie dargcstellt anhand der uridg. Wurzeln

"antreiben",

"suchen" und ihrer Fortsetzer im rigvedischen Sanskrit 353-363
Contact Information of Contributors 365-367
Latin Deverbal Presents in -ii-
MICHIEL DE VAAN (Leiden)
1. The unitary approach
Latin shows a number of presents of the first conjugation which are certainly or probably derived from
verbs, such as cubare, dicare, recuperare, sedare, and the iteratives in -tare and -sare. In the history of
Indo-European studies, one may discern a unitary and an isolationist approach to these presents. The
unitary view is historically the older one. It regards most of the deverbal a-presents as members of a
single category with an inherited suffix *-a-je/o-. In one form or another, this view can be found in
most of the older handbooks, e.g. Brugmann (1916: 160, 164-166), Sommer (1914: 507), Leumann
(1977: 575), and in some of the more recent literature, e.g. Kurylowicz (1956: 303f.), Sihler (1995: 505,
528-530). The main arguments for this interpretation are the presence of deverbal presents with a
similar suffix *-aje- in other branches of Indo-European, such as Balto-Slavic or Germanic, and the
semantic similarity ofthese verbs in Latin. As a representative of the unitary theory, Meillet (1937: 210,
218) distinguishes PIE "aoristic" *-a- from "durative", "iterative" or "reflexive" *-aje-, and notes that
in a number of verbal relics in Slavic, Armenian and Latin, the verbal root shows the zero grade in
front of the suffix *-iiOe)-.
The foundations for a semantically uniform treatment of the a-present were laid by Meillet (1897)
and Vendryes (1910-1911), and elaborated by Barbelenet (1913) and Sjoestedt (1925a, 1925b); compare
also Mignot (1969: 250). Meillet noted that preverbs added to a non-iterative verb tend to render the
verb momentous or perfective: nov/- cogni5vz, sequor- assequor, vorl or-- convortor, calet -- concaluit,
dormii5 -- obdormii5, bibi5 - ebibi5, feri5 - perferi5. The identification of this Latin phenomenon with
aspectual oppositions of the type found in Greek or Slavic has been criticized by van der Heyde ( 1926,
1932-34), Reinhold (1956) and Kravar (1968), among others. Reinhold's alternative distinction (1956:
44) of "Verbalinhalte [ ... ], bei denen der Endpunkt ihres Ablaufs gewohnlich in den Blickpunkt des
Sprechenden tritt, und solche, bei denen dies gewohnlich nicht der Fall ist", foreshadows the verb
classification applied in Functional Grammar, which classifies verbs according to their "state of affairs".
The distinction between terminative and aterminative states of affairs (Pinkster 1990: 214-217) is close
to Reinhold's.
Meillet himself was well aware of the fact that Latin does not show a clear-cut opposition between
"momentous" and "durative" verbs like Slavic, but his use of the terms "perfective" and "imperfective"
has had a negative influence on the reception of his ideas. The topic of the influence of preverbs on
actionality has recently been rediscovered by Latinists such as Haverling (2000, 2003), Romagno
(2003), and Haug (2007). Romagno defines the function of preverbs as changing an unprefixed atelic
verb (faci6, mi5/ior, dormii5) into a telic prefixed one (conficii5, emi5lior, obdormii5).
Meillet also saw that Latin had developed a strategy to provide a perfective compound verb with a
durative counterpart by using the corresponding iterative: aspicere - aspectiire, edTcere - eloquz or
edictiire. One can even discern derivational triplets of imperfective, perfective and imperfective verbs,
such as dTcere- praedlcere- praediciire, or specere- conspicere- conspiciirz (Sjoestedt 1925a: 155),
though the ablaut difference dTc- I die- in the former pair already shows that the three members of the
set did not necessarily arise in this order. As Mcillet himself indicates (1897: 83), this word-
formational principle of Latin is quite comparable to Slavic triplets of the type Russian pis at' 'write'
(impf.) -- podpisat' 'to sign' (pf.) - podpfsyvat' 'to sign' (impf.). Some further pairs of prefixed
durative verbs derived from durative simplex verbs were added by Vendryes ( 1910--1911 ), such as
appelliire to pellere and occupiire to capere. Meillet's observation provides a simple explanation for
316 Michie! de Vaan
the frequency with which a-deverbatives are found in compounds: they were used to create imperfective
counterparts to perfective compound presents. Sjoestedt (1925a: 168) defines imperfectivity in the
Latin verbs as consisting of two ingredients: durativity ("!'action dans son developpement") and atel-
icity ("indetermine", "sans impliquer Ia consideration d'un terme").
2. The isolationist approach
The isolationist approach divides the deverbal a-presents into several subgroups, each of them with a
different phonetic origin of -a-. The category would thus be due to the coincidental merger of several
independent phonetic constellations. This approach was advocated by Steinbauer (1989) and Rix (1995,
1999), and has been followed by Schrijver (1991 ), Meiser ( 1998, 2003 ), de Vaan (2008), Weiss (2009:
400-403), among others. I have now come to the conclusion that the older unitary approach has been
unjustly neglected, and, also, that the isolationist explanations have many shortcomings. Here are the
five main categoties found in modern handbooks, together with my main objections:
a. Intensive/causatives to roots in *-h
2
: doma- < *domHaje- < *domhreie-, tona- < *tonh
2
eie-,
sona- < *suonh
2
eie-, lava-< *louh
3
eie- (Steinbauer 1989, Rix 1999, Meiser 1998: 186-188, 2003).
The principal possibility of some verbs continuing a PIE eie-present after a root in *h
2
cannot be
denied, but it is hard to find a reliable case. For instance, sonclre occurs beside and as a replacement of
sonere, without any visible semantic difference. Thus, sonare may well be a recent formation. For
laviire, Rix (1999: 519) assumes analogical replacement of*lowoje- by *lowaje-, which is ad hoc. See
below on domare.
b. Compounds from nasal presents which generalized the ablaut variant with full-grade suffix,
appella- < *-pel-ne-hr beside pel/ere< thematicized *pelnhre-, also -clinare, -spernari, -sternare, -stinare,
and possibly perfiniis (Steinbauer 1989: 134, Schrijver 1991, 1999, Rix 1995:402, 1999, Meiser 1998:
186-188, 2003).
This hypothesis is insufficiently supported by the formal data, and unattractive from the point of
semantics. The root-final laryngeal is *h
2
in the case of -pellare and -stinare, but *h
3
with -sternare, an
uncertain *h
213
with -.1pernari, while no laryngeal at all occurred in the PIE root of -clmare. Specific
analogies are therefore required to explain the survival of only -na-: for instance, a remake of *sperni5-
l*sperna- into *sperna-/sperna-. The generalization of *-na-to compound verbs and *-na-to simplices
would be based on analogy with the interchange between i and 1 in, for instance, oritur vs. adorltur
(Rix 1995: 403, Meiser 1998: 187). Yet the rise of 1 from *-je/o- in the foqrth conjugation must be
relatively recent (Schrijver 2003), which means that the model invoked for the spread of -nii- arose so
recently that it is doubtful that it could play the role ascribed to it. The question why only the prefixed
verbs take *-na- was solved by Meillet in 1897: prefixed simplices such as appellere and consternere
have perfective semantics which could only be made imperfective by means of a-. Since all old nasal
presents involved belong to the third conjugation, I will treat the compounds in -nii- on a par with the
other dcverbal a-presents.
c.je/o-presents to roots in a final laryngeal: ara- < *araje- < *h
2
erh.;-ie/o-, cala- < *kalaje- < *klhrielo-.
This theory takes for granted the vocalization of laryngeals between obstruents and yod, which is
far from certain (Schrijver 1991: 249f.; pace Rix 1999: 522). The verbs in a are the main evidence for
the vocalization of * H to *a in this position, which makes the assumption that a verb like a rare proves
the vocalization circular. Ferire < *b
11
erH-ie/o- apparently did not vocalize the laryngeal.
d. Denominal verbs to nominal compounds, e.g. educiire to *e-duk-s, occupiire to *ob-kap-s (Monteil
1973: 298, Steinbauer 1989: 139-140, Bammesberger 1996).
I will mention four counterarguments. (l) For most of the a-presents, no cognate compound root
noun is attested. The theory must therefore posit a large number of such unattested compound nouns in
order to explain the attested verbs. (2) The proponents of the denominal theory fail to show that the
meaning of the verbs can be regarded as denominal (that is, essive or factitive). (3) The dcnominal
theory does not explain the distribution of the prefixed a-presents. Many presents only have prefixed
derivatives of the same conjugation as the simplex (baetere 'to go', bibere 'to drink', cadere 'to fall',
etc.), some have prefixed derivatives only of the first conjugation (e.g. -ligiire), while others form both
Latin Deverbal Presents in -a-
317
kinds (e.g. specere - conspicere I conspicarl). I have not come across an explanation for this
distribution in the relevant scholarship. (4) Simplexes such as cubare and iuvare would have been
metanalyzed from earlier compound verbs, such as accubare and adiuvare (e.g. Leumann 1977: 549).
Though there are some simplex verbs which clearly are decompositional, e.g. plicare, there is no such
indication for the majority of them.
e. Frequentatives in -tare. This has been regarded as a denominal category built to the ppp. in -tus
by nearly all scholars, including those of the French school. Sjoestedt (1925a: !55) assumes that the
iteratives in -tare have a denominal origin but somehow "herite de [!]a valeur semantique" of the
imperfective a-presents.
The process by which these alleged denominal presents would have acquired their imperfective and
iterative value is left unexplained, as Weiss (2009: 401) also observes.
My main overall objection to the isolationist approach is that it disregards the semantic properties
of the a-presents involved. Any explanatory theory must take into consideration the possibility that the
a-presents spread primarily for semantic reasons, as is usually the case with productive derivational
suffixes (of course, suffixes can also spread for formal reasons, but that has not been claimed for the a-
presents). I disagree with Steinbauer (1989: 137-138), who quotes an earlier statement by Wissmann
(1932: 204) that the "manifold semantic ramifications" of the deverbal presents are remarkable. On the
contraty, the combination of causative, intransitive or intransitive-durative, iterative and intensive
meanings can easily point to an original type of presents with atelic or imperfective (as meant by
Meillet, Vendryes and Sjoestedt) semantics. Steinbauer's remark (ibidem) that we find all kinds of
ablaut grade in the root syllable of the a-presents, and that this would argue against a common origin of
the suffix, is also beside the point. Such a variety of root forms is the hallmark of a productive suffix,
which will attach to any stem that will have it.
On the basis of formal and semantic criteria, I distinguish three different groups of deverbal a-
presents:
a. Frequentatives in -tare and -sare.
b. Factitives to second-conjugation statives (placare to placere). This type is rarely discussed in this
context, but there are good reasons to include it.
c. Remaining deverbal a-presents.
The following sections will take a new look at the relevant forms with a view to the common formal
and semantic characteristics which they share.
3. The frequentatives in -tiire and -siire
Traditionally, these frequentatives are regarded as a denominal formation derived from the past
participle in *-to- (Lcumann 1977: 547 with references). This explanation is based on their synchronic
status as productive derivatives of past participles, but leaves the frequentative meaning unexplained.
An alternative explanation was put forward by Nussbaum 2007, who argues that the entire frequenta-
tive class arose on the basis of s-desideratives. The development *-T-s- > -ss- in roots in a final dental
led to the merger of the s-stem verb with the past participle in *-Tt- > -ss-. For example, present *kwat-
s- and past participle */cwat-t- both yielded *kwass-. In this way, frequentatives built with a suffix *-s-
a-, such as grassarl 'to go round, approach' if from *grad-s-a- (to gradior), pens are if from *pend-s-a-,
etc. (to pendo), could be metanalyzed as belonging to the past participles *grasso-, *pensso-, etc.
Subsequently, the use of the stem form seen in the past participle would have become obligatory in all
frequentatives, yielding verbs such as captare, iactare, gestare.
The main evidence for the original formational pattem of root plus *-s-a- posited by Nussbaum is
provided by relic forms which are not synchronically linked to a past participle. For instance, taxare 'to
assess (the worth of)' to tango 'to touch' can historically be regarded as a relic form since it was not
replaced by *tactare. Other relic verbs are axare 'to name' (to aiio), cassiire 'to totter' (to cado),
fraxare 'to go the watchman's rounds' (if to jrequens), rapsiire 'to hurry along' (to rapio), rixarf 'to
struggle' (rixa 'quarrel'), and vexare 'to agitate, damage' (to veho). There is no extant nominal basis
for any of these verbs except for rixarf and possibly vexare (vexus). I therefore agree with Nussbaum
318 Michie! de Vaan
that we are probably dealing with a-presents derived from Proto-Italics-presents of the form *kad-s-,
*rap-s-, *tag-s-, *vex-s-, etc. In fact, some of theses-presents have been preserved ins-future forms in
Old Latin, such as surrepsit and taxis. Due to the disappearance of the athematic indicative of the s-
present in Latin- as opposed to its preservation as an s-future in Sabellic --the frequentatives in *-sa-
were reinterpreted as independent a-presents, and were dissociated flom their root: for instance, vexare
was dissociated from vehere.
Nussbaum 2007 assumes that the a-conjugation in these frequentatives ultimately arose in denominal
presents, as one might suspect for rixarl. The missing link would be an adjective in *-so-, to which the
presents in *-s-a- could be the regular denominatives. Such a formational pattern is unobjectionable in
general, but I find it difficult to accept in this case. A minor problem is that adjectives in *-so- are only
(indirectly) attested for rixarf and vexare. The main problem is that a denominal origin does not explain
the imperfective and iterative meaning of these verbs. It seems more likely that they contain deverbal
*-a- which was added to Proto-Italics-presents of the type *rap-s-.
The main semantic characteristic of Latin frequentatives is their repetitive and/or atelic state of
affairs (Bodelet 1913: 175-213, Sjoestedt 1925ab). This aspect clearly applies to the relic verbs in -sa-,
all of which express atelic movement, that is, movement without a natural goal: cassare 'totter', errare
'wander',fraxare 'go the watchman's rounds', grassar'i 'go around', quassare 'shake vehemently',
pensare 'weigh up, balance', rapsare 'hurry along', rixari 'struggle', taxare 'assess', versarf 'keep
turning round', vexare 'agitate, damage' (axare is attested without context). Many atelic movements
are repetitive, such as 'totter', 'err', 'shake', 'turn', 'agitate', which would explain why -sare and
hence -tare became productive for iteratives and frequentatives in the way that Nussbaum envisaged;
compare Kulikov 2008: 326-328 for the connection between atelicity and frequentativity. Since *S-
had a conative (traditionally "desiderative") meaning, it must be *-a- which provided the atelic semantics.
The rise of the type in *-s-a- must be relatively old. The suffix vowel cannot be equated with the a-
subjunctive in its synchronic meaning since it would have competed with the subjunctive in *-s-f-(<
optative). Also, the atelic movement would be difficult to explain in this way. In fact, atelicity is also
the characteristic feature of deverbal a-presents of the type cubare, dicare, as we shall see in section 5.
It is therefore tempting to assume that the frequentatives in -sare actually contain the same atelic
present suffix *-a-je/o-. In that case, the frequentatives are historically just a coincidental subclass of
the Proto-Italic deverbal presents in *-aje- ( deverbal since *-s- was a verbal suffix).
The same suffix *-s-aje- may also be contained in three Latin verbs in -erare. In my view, the Italic
s-suffix still possessed its Proto-Indo-European suffixal ablaut *-s- : *-es-, as is clear from the Sabellic
evidence and from the Latin perfect subjunctive in -er-; see Pedersen 1921 and de Vaan forthcoming.
Depending on the ablaut fom1 of the suffix, a derived a-present would therefore have the form *-s-aje-
or *-es-iije-.
Lat. recuperare 'to take back, regain', recuperator 'assessor', can reflect *kap-es-a- 'to try to grab',
with the same stem *kap-es- which is found in the present capessere. The conative meaning is clearly
present in recuperare, which may thus be interpreted as the atelic counterpart of telic recipere 'to
admit, acquire'.
A similar case is lamberare 'to beat', attested in Plautus, Pseudo/us 743: meo me ludo lamberas;
and in Paulus ex Festo: lamberat 'scindit, laniat'. It may reflect *lamb-es-a- 'to try to lick' to lambere
'to lick'. The semantic development of 'lick' to 'beat' is comparable to that found in English lick 'to
lick; beat'; note the conative meaning included in the figurative use.
Tolerare 'to endure' can hardly be explained as denominal, since Latin has no noun *telos-. The
atelic meaning renders it likely that the verb is a deverbal a-present. Two theoretical options are a verb
stem *telas-a- < *telhrs- with vocalization of the laryngeal (thus Nussbaum 2007), or *te/-es-a- 'try to
bear> 'endure' with *tel- as in Latin -tulat and the Umbrian future en-telust. Since the latter hypothesis
is based on a known stem variant, it may be preferred.
Latin Dcverbal Presents in ii- 319
4. Factitives to second-conjugation statives
A handbook example of this class is placare 'to appease' to placere 'to please'. At first sight, one
might regard the factitives as a subclass of the denominal factitives of the type novare, but a suitable
nominal basis is only attested for sedare (viz. the noun sedes 'seat'), for which a verbal origin is
equally available in the perfect sedf 'I sat'. It seems more likely that these presents are deverbal, -a-
causing the factitive meaning.
We have already hypothesized that *-iije- can change the verb's actionality from telic to atelic, and
also that the frequentative meaning of the iteratives in -sa- and -ta- has its source in verbs of atelic or
repetitive movement. Cross-linguistically it is not uncommon for intransitive verbs to become transitive
or causative when provided with a morpheme expressing iterativity or atelicity; for a general
introduction to this topic, cf. Kulikov 1999. Four or five of the six verbs in this class are in fact verbs of
movement. A good example is sed are 'to restrain, cause to lie down', literally 'to make someone sit'. It
can be viewed as a derivative, not of stative sed eo, but of its resultative perfect sed! 'I sat'. Suffixation
as *sed-a- with the suffix of atelic movement would cause the change from stative 'to sit' to dynamic
'to sit down (in various directions, in various places)', while the concomitant causative meaning 'to sit
(someone) down forcefully, restrain' is explained by the increase in effectivity and transitivity which
often accompanies atelic movement (Kulikov 1999: 22, 26).
This explanation implies that the link between factivitive a-presents and stative e-presents may be
accidental: the derivational base was stative, and could be furnished with *-eje- (expressing stativity) or
with *-aje- (atelic movement> [activity). I will discuss the evidence in alphabetical order of the root
verb.
4.1 Creiire 'to make grow'. The classical meaning 'to engender' is more recent; all six attestations in Plautus refer
to 'causing satisfaction I troubles I delay I opportunity' or 'committing a crime'. Thus, while the semantics are
factitive, they can in most cases be regarded as atelic. This present belongs to crescere 'to be born; increase', pf.
crevl, from a Proto-Italic root *kre- 'to grow'< PIE *k(wJrehr. To the same root, creiire can represent a derivative
*kre-iije- with the suffix of atelic movement.
4.2 Deliciire 'to reveal, disclose' to liquere 'to be clear'. Whereas Old Latin consistently has de/ici5 with the
meaning 'to reveal, disclose', from Varro on we find deli quo 'to strain (a liquid)'. Thus, delici5 with its derived
meaning must be older, whereas qu was apparently restored in deliqui5 on the basis of liquei5.
Since the primary and etymological meaning of stative liquere is 'to be liquid', an explanation for delici5 based
on the specific meaning of atelic movement of *-iije- is quite conceivable: 'to make liquid' = 'to make flow', and
'flowing' is typically atelic (in many directions, or all the time). There is also a third conjugation verb tiquor 'to
become liquid, dissolve; flow, run' < *wleikw-. Thus, although deliciire synchronically looks like a derivative from
liquere, it may actually have been derived from the zero grade of an older athematic stem *wleikw- I *wlikw- 'to
be( come) liquid'.
4.3 Pri5mulgiire 'to make widely known' to mulgere 'to milk' to PIE *h
2
mlg- 'to milk'? This etymology is
uncertain. Mu/gere may continue a PIE causative *h
2
mo/g-eie- 'to milk' or a stative verb *h
2
mlg-ehr 'to give
milk'. Pri5mulgiire could be interpreted, with some imagination, as 'to milk forth', whence metaphorical 'to bring
out, spread, make known'. The verb is only attested from Cicero onwards, which seems to point to its recent
formation; but in view of the meaning, that is hardly likely (unless it was a loanword of some sort). In support of
the possibility of an earlier derivation of pri5mulgiire from *h
2
mlg-, note that mulctra I -um 'milking pail' is
another old derivative independent of mulgere.
Of course, it remains possible that the connection of promulgiire with mu/gei5 is only apparent. Forssman
(2002: 180-181) has proposed to derive pri5mulgiire from *pri5-morigiire 'to postpone', dissin;ilated to ~ p r i
molgiire whence pr6mu/gare. However, this hardly seems a more likely alternative than the solution proposed here.
4.4 P/iiciire 'to make favourably disposed, appease' to placere 'to please'. The meaning of p/iiciire is not only
transitive but also appears to be telic, and the verb already occurs once in the perfect in Plautus. An interpretation
as a verb of atelic motion is possible if the root *plak- originally meant 'flat' (to PIE *plehr 'flat') and p/aceo 'to
be even, not to make a difference'. The factitive would then be 'to make even, flatten'. Scbrijver (1991: 181)
suggests that p/aciire has the original PIE root ablaut whereas placere would have a secondary shmt a by analogy
with other e-verbs, sucb as manere and patere. While not completely impossible, such an analogy seems unlikely,
since Latin tolerates root ablaut in quite a number of verbs, and since a stem *p/iicere, had it existed, would have
been supported by pliiciire. By Schrijver's own rules (*CRHTC >Proto-Italic *CRaTC, Schrijvcr !991, 2006), a
preform *p/Hk-C- would have yielded *plak-C-, which might be the origin of the ablaut form inplacere.
320 Michie! de Vaan
4.5 Irrigare 'to make wet, irrigate' to rigere 'be stiff'. The simplex rigare is nol attested until Lucretius, and was
probably backformed to irrigare. Due to the vowel weakening in unstressed syllable, there arc two verbal roots to
which irrigare could belong: *reg- 'to lead' as in regere 'to lead', and *rig- 'to stretch' as in rigere 'to be stiff',
rigidus 'stiff' and Proto-Celtic *rig- 'to lie' (de Vaan 2008: 523). In either case, irrigare could be a verb of ale lie
movement in *-aje-.
4.6 Sedare 'to restrain, calm down, put to rest', consedare 'to check, stop' (Cato) to sedi.!re 'to sit'. In Plautus, the
passages with sedare favour an atelic interpretation, for instance, Bacchides 108: Sequere hac igitur me intro in
lectum, ut sedes /assitudinem 'Come inside and shelter, then, so as to case away your weariness'. Walde-Hofmann
(II: 508) explain sedare as a denominal verb 'to put into the seat' to sedes 'seat', from a root noun which may be of
Indo-European origin. While possible, there arc no positive indications that this assumption would be correct. A
good alternative is a derivation from the verbal stem *sed- as attested in the Latin perfect sed/ 'I have sat down, am
sitting'< *se-zd- to both sedere 'to sit' and sldere 'to sit down' (Meiser 2003: 203f.). The original meaning of sed-
a- can be interpreted as 'to sit down' > 'cause to sit down, put to rest'. Consi.!dare can either be regarded as a
pcrfectivizing compound in con-, or as an a-derivative of the perfect c6nsedT.
5. The remaining a-presents
The previous two sections have shown that an analysis of *-aje- as originally referring to atelic
movement is able to explain the repetitive meaning of the iteratives and the factitive meaning of the
factitives. Below I will discuss the evidence of the remaining deverbal a-presents. As indicative for
their appurtenance to this class I regard any one of the following clues:
I. The present has atelic semantics.
2. The present is attested together with other present stems of the same root, either in Latin or in
Sabellic.
3. The present has a strong perfect and/or past participle, and/or other devcrbal derivatives which
point to a non-first-conjugation stem.
4. The present cannot be easily explained as a denominal formation.
5. There is evidence for verbal character of the root outside Italic.
It will be clear that there are ambiguous cases, for which a decision between a deverbal or a denominal
a-present is hard to make. In order to show that the valency of the basic verb and the status of the
derived verb as a simplex or a compound arc irrelevant to the possibility of taking the suffix *-aje-, the
evidence is divided in four groups according to those two criteria.
5 .I Simplexes from an intransitive base
calare 'to announce, summon', calator 'personal attendant, servant'. The main usage of calare was for 'summoning'
a group of officials, which can be regarded as a repeated action, viz. calling out the name of each of the persons
which was summoned to attend the gathering. The same verbal root is also reflected in Latin concilium 'debate,
council' < *kom + *kV/io-, and possibly by ca/umnia 'false accusation' < *kalomno-. In Umbrian, the third singular
imperative kal'etu, karitu, carsitu 'must order, must call' (Garcia Castillero 2000: 271) is generally assumed to
reflect a present stem *kal-e-. The co-occurrence of Sabellic *kale- and Latin *kala- has not been satisfactorily
explained. The least unattractive solution is Schrijver' s conjecture of an original ablauting present * kelh
1
ti,
*klh
1
enti 'to call'> *kelati, *kalenti, whence Latin generalized a mixed stem *kal-a- which then switched to the a-
conjugation. Yet the last two changes are ad hoc and without certain parallels, and they still leave Umbrian -e-
unexplained. Garcia Castillcro (2000: 212) proposes that Sabellic may have had simple primary presents next to
iteratives in *-e- (as in kumaltu next to maletu). If this is true, we may posit a primary verb *kal-(e/o-) 'to call'
(possibly reflected in calumnia) next to which Latin made an atelic counterpart in *-a- and Umbrian one in -e-.
crepare 'to make a sharp loud noise, rattle, crack'. Clearly an atelic verb of sound, with durative and probably
iterative semantics. IfVarro's percrepis 'you sound throughout' is based on an earlier present *crepere, and is not
a backformation to crepuf, then crepare can be interpreted as the repetitive counterpart to *crepere. The noun
crepundia 'child's rattle' looks like the derivative of a gerundive of the same *crepere.
The root has sometimes been reconstructed with a final laryngeal (Rix 1999: 519, LIV
2
373) in order to explain
Latin -a- as indirectly the result of laryngeal vocalization and because of Rigvcdic but the latter can
easily be a recent i,l'-aorist. Hittite karp- 'to be angry' cannot tell us whether the root had a final laryngeal;
Klockhorst (2008: 452) reconstructs PIE *kr(e)p-.
Latin Deverbal Presents in -a-
321
cubare 'to be lying', accubare 'to be lying at, to recline at table', incubare 'to be lying on, brood, cover', recubare
'to lean back, rest'. The stative meaning of * kub-a- is proven to be of Proto-Italic date by Faliscan cupat, cupant,
SPic. qupat, Pacl. incubat, Marr. cibat 'is, arc lying'. The cognate dynamic present is Latin -cumbere 'to lie down':
accumbere 'to recline at table', incumbere 'to lie on', etc.
This verb is exceptional among the a-presents in that * kub- had a stative meaning. Also, the suffix -a- does not
seem to alter the meaning, unlike in the category of sedare, where original stative verbs turn transitive. For these
reasons, one may look for an alternative source of its a. It may have been adopted from its antonym stare 'to stand'
< *staje-. Since 'to stand' is formed by a different stem sta'i/e- < *sta-ehr in Sabellic (Garcia Castillero 2000:
359ff.), Rix and Meiser suppose that cubiire reflects a stative formation in *kuba-e-. But Sabcllie and Faliscan
/kubiit/ cannot reflect *kuba-e-, since the phonetic development apparently was different from that in disyllabic
sta-T/e ... The root-final laryngeal in alleged PIE *keubhr (LIV
2
360) is only reconstructed on the strength of the
Italic a-suffix, which makes the explanation of cubiire from *keubhr circular.
labare 'to stand unsteadily, waver' to /abf 'to glide, slip', /abundus 'gliding'. Semantically, labiire can be a
frequentativc to liibf. The compound labefacere 'to make unsteady, bring to fall' is the only factitive in
which is not derived from a stative second conjugation verb or from related adjectival forms; compare Balles 2004:
218-226. Semantically, is a factitive to labare. This renders it likely that both labare and labefacere
contain the ablaut variant *lab- of the root of /abf, thus possibly pointing to an earlier ablaut */Zib-: *lab-.
meare 'to proceed, traverse' has a frequentative counterpart commetare 'to go constantly' which is not certainly
explained. It could go back to a past participle *mejeto-, which would imply that the root *mei- possessed verbal
derivatives in Proto-Italic, and hence meare could be a deverbal a-present to the root *mei-. The atclic semantics of
meare would support that assumption. A nominal source is not present in Italic, the only nominal derivative
attested being the 1-stem *mit- in semita 'side-path, track, pavement' and trames 'footpath, track, course'.
According to an old theory, meare might be derived from the root *mehr 'to measure', which is represented in
Latin by the verb metTrf, a denominal derivative of *mehrti- 'measurement'. Besides 'to measure', metfrf is also
used metaphorically for 'covering' a distance and 'passing' across a certain background. If these meanings were
already present in *mehr at an earlier stage, meare could be interpreted as Proto-Italic *me-aje- 'to traverse'.
micare 'to quiver, dart, flash', emicare 'to dash out, jump forth'. 'Quiver, flash' is clearly a repetitive action. In
view of the absence of old nominal formations, micare is probably deverbal. Cognate verbs occur in Celtic (Old
Irish dimicnigid 'despises', Welsh ed-mygu 'admire') from *mik-n- (Matasovic 2009: 271) and in Slavic *mikati
'move abruptly, blink' < *meik-. Ll V
2
hesitantly reconstructs the root as *meikhr 'to flash', but the final laryngeal
is only based on Latin -are.
so/arf 'to give solace, comfort', consoliirT 'to comfort, offer consolation or compensation'. 'Comforting' is
typically an atelic action. The verb has been connected with Gr. i:>cucrKO!!at, 'iAct!lat 'to appease', from a PIE root
*selhrl*slhr 'to be( come) merciful'. Since the meaning of the PIE root is assumed to have been intransitive, Latin
solar/ has apparently switched to being transitive. This is reminiscent of the explanation of sedare as an a-present
to the perfect sed/; it would presuppose a perfect *sol- 'having become merciful= to be merciful', but the rise of o
in the perfect is hard to explain. Sehrijvcr (1991: 126-128) interprets so lor as denominal to a root noun *sol-.
soniire 'to sound' to sonere 'id.'. Both verbs have the same meaning, but in the course of time, sonare has ousted
sonere. Similarly, we t1nd resonare 'to resound' besides resoncre. The semantics are atelic and dmative. It cannot
be excluded that sonare is a denominative verb to sonus 'sound' (Schrijver 1991: 395), but in view of the existence
of sonere, this view is not compelling. It is equally possible that sonere continues PIE *suenH- (we> o in Latin),
beside which an atelic or iterative a-present arose at some stage. The meaning ofUmbrian sonitu, sunitu [2s.ipv.II],
used in a curse, is too unclear to warrant that it is cognate with sonere.
tonare 'to thunder', cantonal 'it thunders violently' (Piautus) to *!onere in Varro (tibi typana ... tonimus). As with
sonare versus sonere, it seems that the a-present has ousted an older thematic present. Only Sanskrit shows verbal
forms of this root, which point to a root structure *(:,)tenH-. A denominal origin of tonare seems unlikely since
tonus does not mean 'thunder' but 'pitch, tone, note', and is only attested in Imperial Latin. If !onere and tonare
are not denominal, their o may have been adopted analogically tlom sonere, or it must continue a real PIE *o.
volare 'to fly'. A verb of atelic movement in most Plautine occurrences. Of course, one may also think of the
repetitive action of a bird's wings. The etymological appurtenance of this verb is disputed, but I find the
connection with *gwelhr 'to throw' attractive. This root is reflected in Celtic *ad-bat- 'to die', Greek Ba:>c:>cw 'to
throw' and maybe in Skt. ga/- 'to raise the arm'. Pit. *gweza- would yield Latin val-; it is uncertain whether *g"'ol-
a- would also yield vola-, or whether unrounding after *wo- would have yielded *wa- before the I could round the *e.
The adjective volucer 'able to t1y, winged' points to *gwe/ulo-tli-. Since a stem *g"'el-u- is unknown (the
connection with Skt. garur;la- 'mythical bird', supposedly from *garutra-, cited by Weiss [2009: 318] is very
hypothetical since *g"'olutlo- would yield Skt. *garutra-, whereas *gwelutlo- would yield ), and since
322 Michie! de Vaan
alacris 'nimble' < *ala-tli- 'moving aimlessly' (Weiss 2009: 318) shows that -tli- was suffixed directly to the
verbal stem, it seems more likely that vu/ucer is based on a verbal stem *g"'el-e/o-.
5.2 Compounds from an intransitive base
amhu/iire 'to walk around, go for a walk' < *amb-aliire. A probable cognate is Umbrian amb-oltu 'must go
around'< *-ala-tad (Meiser 1986: 270) <PIE *h
2
elhr(e/o-); cf. Garcia Castillero 2000: 257f. Since the Latin and
Umbrian verh arc only attested with *ambi-, it has been supposed that the a-conjugation is automatic because of
the compound, whereas Schrijver (I 991: 400) posits a present *-a/a)e/o- in Latin versus *ala- in Umbrian. But if
we adopt Mcillcl's semantic analysis, ambu/are may simply be the alelic counterpart 'walk up and down' oftelic
Proto-Italic *amb-al(e)- 'walk around' attested in Umbrian. [f alacer 'lively, active' indeed reflects *ala-tli-, the
present to which ambulare belongs was *ala- from athematic *h
2
elhr.
acc/Tnare 'to lay down, rest on, lean towards', dec/Tnare 'to deflect, divert', indinc!re 'to cause to lean, bend
downwards', rec/lnare 'to cause to lie back' to *c/inere 'to lean, bend' rrom PIE * ki-n(e)-i-. Apparently, i was
introduced into the root in Proto-Italic, yielding a stem *klin-e/o- or *klin-i-. Since PTE *kli- had no final laryngeal,
Praust (2000: 141) prefers to disconnect -dinare from other Latin presents based on PrE *-n-H-, such as pel/ere,
and he suggests that it may be a denominal verb to the same noun as found in Gr. KAtVll, OHG (h)/ina 'rail, back'.
However, the semantics of -clfnare are hard to interpret in a factitive or essive sense from a noun 'rail, back,
support'. Hence, we may consider -cllnare to have the devcrbal suffix *-aje- of atclic movement, which, added to a
stativc verb 'to lean', yielded a dynamic and transitive meaning 'to cause to lean, bend', as with the type sedare.
Note also that -cllnare contains the full--grade root *klein-, just like the other a-derivatives from nasal presents.
5.3 Simplexes from a transitive base
amare 'to love'. This durative verb has a Sabellic cognate in Marrucinian 3pl.pf.ind. amatens, which confirms a
present stem *ama-. Within Latin, amor 'love', amicus 'friend', and probably also amp/us 'large, big'< *am-lo--
point to a root *am-. This is often connected with Skt. am
1
- 'to take hold of, ama- 'attacking power' and Gr.
OflVUflL 'to swear' from PIE *h
2
(e)mh,-. Rix (I 999: 523) explains amiire from *amajelo-, with vocalization of* H
in PTE *h
2
(e)mhrielo-. As argued above, the assumed vocalization of *His uncertain. A viable alternative is to
derive amiire with the suffix *-aje- from the synchronic Proto-Italic root *am- 'to lake hold of.
ariire 'to plough'. Typically used atelically as 'to be ploughing'. The evidence of other languages enables us to
reconstruct a present *h
2
crhrielo-, which e.g. Rix (1999: 523) and LIV
2
assume to have yielded Pit. *araje/o-,
whence with contraction Latin ara-. The same criticism about the vocalization of *Has above applies here too.
Celtic does not show vocalization of the laryngeal: Mlr. airim, W. arddu 'to plough' <Proto-Celtic *ar)o-. An
alternative solution as *ar- plus the suffix of atelic movement is worth considering. Note that 'ploughing' is a
typical repetitive action of 'breaking' the ground again and again (cf. Hittite harra-
1
I harr- 'to grind, crush' <
*h
2
orhT, see Kloekhorst 2008: 300f.), or of going back and forth across the land with a team of oxen.
celclre 'to conceal from view, keep secret' to occulere 'to hide, conceal'< *ob-kel-e/o-. The root form *kel- could
arise in the root aorist or, with early simplification of the reduplication, in the perfect *ke-kl- >*kef- 'is hidden'. In
that case, the semantic shift is similar to sed- 'to sit' > si?dare. Many scholars assume that ce/are is denominal to a
noun withe-grade, cf. Steinbauer (1989: 142) and Schrijver (1991: 124). The obvious candidate for such a noun
would be a root noun *kel- I *kef-, but that is not attested.
dicare 'to assign, dedicate; indicate', abdicare 'to deny, renounce', dediciire 'to declare', indicare 'to make known,
reveal, betray, fix the selling-price', praediciire 'to make known, proclaim, announce', all from *dik-a-, to dicere
'to say' < *deik-e/o-. The meaning of dicarc is clearly atclic (Plautus, Bacchides 995: ubi lube/ recita: aurium
operam tibi dico 'When you're ready, read: you may command my ears'), and the same goes for most compounds.
It is possible that a repetitive component is involved in the meaning of these verbs. In Oscan, the same formation is
probably found in the 3sg. preterite dadikatted 'has offered' from the preverb da(t)- and a present stem *dik-c/-
with short *i; Garcia Castillero (2000: 280) dismisses Untermann's suggestion that the Oscan form might be a
loanword from Latin.
It has been assumed that diciire is a backformation to the preverbial compounds, but this seems questionable.
The most important counterargument is a semantic one: whereas dicare has a general and atelic meaning, dicere
'to say' is strictly tclic; and the same difference holds, for instance, between indicare 'to make known, disclose'
and indicere 'to give formal notice of, proclaim (a public meeting, a festival, etc.), declare (war), impose (a task or
punishment)'. Steinbauer (1989: 140) strongly pleads for a derivation of indicare !rom the noun index, attested
since Accius. This would indeed explain the short root vowel directly, and the meaning 'to be an indicator' can
easily lead to the attested meanings of indiciire. But the short root vowel can equally well be explained if *-iije-
Latin Dcverbal Presents in -a-
323
was sufilxcd when Proto-Italic *deik-!*dik- was still an athematic present. Steinbauer does not discuss the Sabellic
cognates, which would require a much earlier date for the denominative.
It may seem much harder to deny a denominal origin for compounds with a nominal first member, such as
iiidicare 'to judge' to iiidex 'judge', vindicare 'to lay claim to' to vindex 'surety', and Oscan medicatud 'judged'
(abl.sg.) to meddik- 'judge'. Still, all three verbs refer to the profession of the referent, which may be interpreted
as atelic activity, so that an explanation as compounds with dcvcrbal dicare is equally possible. In that case, index,
iiidex, meddiks were backformed to the compound verbs.
do/are 'to hew or chop into shape; batter'. The semantics arc clearly repetitive. Rix ( 1999: 527) derives do/eo 'to
hurt, be in pain' from a causative/iterative *dolhreie-, whereas do/are would reflect intensive *delhrie- > *delaje-
> *dolaje- or *dlhrie- > *dalaje, in both variants with vocalization of the laryngeal to *a. Schrijver (1991: 215)
suggests a denominal origin from a noun *dola-. A good alternative seems to me that dolare contains the suffix
*-aje- of atelic movement.
domare 'to subdue' to PIE *demhr 'to tame, constrain'. Because of its transitive meaning, domare is usually
derived from a PIE causative *domhreie- >Pit. *domaje-, but a closer inspection of its usage in Old Latin renders
an imperfective interpretation quite likely, as was seen by Rix (1999: 521 fn. 36). The Oxford Latin Dictionary
lists the meanings 'domesticate', 'habituate to', 'reduce to subservience, gain control over', 'bring under control',
'reduce to a more amenable form, bring under cultivation', all of which stress the process of bringing under control.
Hence domare may reflect a Proto-Italic present in *-aje- with atclic semantics.
forare 'to bore (through), pierce'. 'Boring' is typically atclic and repetitive. The a-grade of the root cannot be
explained from earlier *e or *a, as in the case of doliire, so that we must find a preform with PIE *o. Lith. harti 'to
accuse, scold', OCS brati (.Ye), bor}Q 'to fight' reflect a PIE a-grade present *bhorH-. If this survived as *for- into
Proto-Italic,.forare could reflect the addition of the suffix *-iije- ofatelic movement.
iuvare 'to help, assist' to iuvere 'to help' (in Accius) and hlcundus 'pleasant'< The present iuvere <
*(i)juwe/o- < *(i)jow-e/o- is thought to reflect a reduplicated formation *Hi-HeuH-e/o-. As in the case of sonare
and tonare, the synchronic evidence suggests that the present in -are has replaced the original third-conjugation
present.
lavare 'to wash, wash oneself; bathe' to Iavere 'to wash (something or someone)'. The a-present is clurative and
atclic, whereas Iavere and its compounds abluere, dT/uere, etc. arc telic. The a in lav- is probably clue to
unrouncling of *o in *low-e/a .. , as explained by Vine (2006: 239). Tn Umbrian, a present *!awe/a- or */owe/o- is
represented by vutu [3s.ipv.TI] and anderuomu (Garcia Castillero 2000: 381 ).
Lavare is traced back to *lawa-e- by Schrijver (1991: 397), but whether *-ae- would regularly yield long a is
uncertain (I would rather expect e), and furthermore 'bathe' is not stative but clurativc. Rix (1999: 5 19) and Meiser
(1998: 187) assume a phonetic development of *louhreie- to *lowoje-, which was then reformed to */owaje-. An
alternative explanation would be to assume that the atelic a-suffix was added to the stem *law- of Iavere.
/ibare 'to pour a libation of; take a small sample of, nibble', de/Tbiire 'to remove a small portion of'. The semantics
of the verb arc mainly atelic, and in fact 'to libate' for the gods is usually a repetitive action. Libare has been
regarded as clenominal to !!bum 'sacrificial cake' (Steinbauer 1989: 120-122), but I doubt that the wide range of
meanings of /Tbare can all be explained on the basis of llbum. Therefore, I would suggest that the original present
*leibh-(e/o-) 'to anoint, sacrifice to' (cognate to Greek aA,cicpco 'to anoint') yielded in Proto-Italic, on the one hand,
a derived noun *leifo-, and, on the other hand, an atclic present* leila-.
mitat vs. mittere. The form mitat (/mitat/ or /mitat/) is attested in the Archaic Latin Duenas inscription (iouesat
deiuos qoi med mitat 'Iura! cleos qui me donal') and on the Tibur base (hoi med mitat Kavios 'Hie me clonal
Gaius'). Whereas the latter certainly contains an indicative form, Duenas' mitat is usually also interpreted as
indicative, but a subjunctive cannot be completely excluded. Vine ( 1999: 207) has explained mitat as a
denominative verb *mila- 'to exchange' to a past participle *mi'to- 'exchanged' to the PIE root *h
2
mei. More
traditionally, mitat is taken as a form of mittere 'to release, send, throw', which may go back to *mTt-e!o- to PIE
*meithr 'to exhange' by the littera rule. If it does belong to mittere, mitat could be a verb of atelic movement in
*-aje-: *mit-aje- 'bestows'. The zero grade in !mit-a-t/ as opposed to the full grade in *meit-e/o- corresponds with
the distribution in dicare versus dicere. Rix (1999: 521) offers a different explanation: an old iterative *moithreie-
> *moitaje- would have replaced the root syllable *moil- by *mit- on the model of the conjectured nasal present
*mitn-!*mitan-. This seems too complicated to be convincing.
(male) mulcare 'to handle roughly, beat up, discomfit, mutilate'. In Plautus and Terence, the verb is mostly found
together with the adverb male 'badly', which suggests that the original meaning of atelic *m(o)lk-clje- was
connected with mulcere 'to stroke, caress, soothe'< *m(o)lk-eie-. A nominal origin for mulcare, as proposed by de
Vaan (2008: 393), is not supported by the semantics of the verb.
324 Michie! de Vaan
pardre 'to prepare, make preparations; make an effort to obtain; acquire; furnish, provide'. The semantics go from
intransitive ('to prepare for') to transitive ('to obtain'), and from atelic ('to be preparing') to tclic, especially when
used in the perfect. The meaning shows that parare must be connected with parere, pari a 'to give birth to, bear'; it
may thus be a deverbal a-derivative of the antevocalic stem *par- < *prhr 'to provide'. The atelic compounds
imperare 'to demand the production or payment of, order, mobilize (the army), to give orders, be in command' and
properdre 'to act with haste, hurry, to bustle with activity; to hurry on' can be explained from *in-parare 'to bring
on> demand' and *pro-parare 'to bring forth' with word-internal vowel weakening; thus Panagl!Lindner (1995:
173f.). While they probably regard the a-conjugation as the automatic result of the 'intensive' derivation to parere,
I prefer to see these verbs as semantically justified counterparts to *in- and *pro-parere. The compounds with
unweakened -par-, such as appardre 'to prepare, provide' and comparare 'to prepare, buy, obtain', must be of a
more recent date. They are clearly derived from parare, not from parere.
runcare 'to grub up (plants), weed'. The action of 'grubbing up' is repetitive. One might conjecture that runcare
was formed to a noun *runko- '(the action of) weeding', but there is no trace of such a noun. The basis is more
likely to have been a nasal present *Hru-n-k-, which may be compared with Skt. luncati 'to pluck', Latvian rfiket
and Greek 6pucmw, 6punw 'to dig, scrape, bury' < *h
3
reuk- (but *h;rug"-ie/o- is reconstructed for Greek by
Beekes 20 I 0: 1113). Whereas the original nasal present was lost in Latin, the derived present *runk-aje- of atelic
movement was preserved.
secdre 'to cut in pieces, slice, cut a portion from, mow, castrate'. Typically a repetitive action, which can, in some
contexts, develop into telic 'to cut off'. There is no indication that the present has a denominal origin, as Garcia
Castillero (2000: 338) rightly holds against Steinbauer 1989: 132-134. The latter scholar does, however, cmrectly
point out that other derivatives of the root *sek- betray an earlier present *sek-(e/o-), such as signum 'mark, sign'
and Varro's secula 'sickle'. Similarly, the perfect and past participle also presuppose a root *sek-. In Umbrian, the
same present formation *sek-a- occurs in prusekatu [3s.ipv.IJ] 'must cut out', whereas the ppp. [n.pl.]
'cut out' and -es [f.] 'uncut' reflect *selc-eto-. Pace Rix (1999: 526) and LJV
2
, I think that *-eto- in the past
participle of several Umbrian a-presents (-se9eto-, oseto, muieto, pesetom) does not prove an original root-final
laryngeal, but may be due to the (limited) productivity of the perfect in -ed in Sabellic. Thus, Umbrian * seka- next
to *seketo- is very similar to the Latin situation, where a number of presents in -a-, among them secare, take a
strong perfect in -ul, -itus. This confirms that, in Proto-Italic, *-i'ije- was a productive present suffix.
The suggestion, made by Rix (1999) and adopted by LJV
2
, that Latin -i'i- could reflect vocalization of *h
1
in
front of a suffix *)elo-, is rendered irrelevant by the semantic and formal analysis proposed here. Still, for this root,
a PIE root-final laryngeal may have to be reconstructed. Kloekhorst 2008: 696 argues that the PIE root must have
been *skhr in order to explain Hittite -kk- in sakk-/akk- 'to know'. Latin scire may continue an original i-prescnt
which underwent the same semantic shift as the Hittite verb.
venarT 'to hunt, chase'. Typically a verb of ate lie movement. Within Italic, *wend- is the only verbal stem
representing this root. Another derivative is venia 'favour, permission', cognate with Oir.fine 'family', OHG wini
'friend' and wunnia 'pleasure'. The root may be reconstructed as PIE *uenhr (thus LTV
2
) or *h
113
uenhr
(Kloekhorst 2008: 999f.) 'to desire', on the basis of forms such as Hittite yen-zi I uyan- 'to copulate' and Skt.
vanate 'to love', vama- 'sweet, precious'. The long e in the Latin present, if of PIE date, could have originated in a
root noun or in a monosyllabic verb form, such as the 2sg. or 3sg. *h
113
uenhrslt of a root aorist injunctive.
vetare 'to forbid'. The original root vocalism could be Proto-Italic *wot- or *wet-; cf. de Vaan 2008: 672. Rix
(1999: 520) reconstmcts a present *vetere which yielded the perfect and ppp. in vet-, but he supposes that vetare
continues an old causative *uothreie-. Yet there is no semantic support for the assumption that vetdre was an
original causative. A cognate present *wet-e/o- 'to say' is attested in Celtic, where Old Irish aspena 'testifies'
could go back to *eks-uet-nd-; the nasal present suggests a root in *-h
2
.
vorare 'to swallow, devour'. It has atelic semantics and may thus be opposed to telic devorare 'to swallow (up)'.
Vorare could rei1ect earlier *g"'er-i'ije- or *gw or-aje- from the PIE root *gwrhr 'to devour'. It is explained as a
factitive denominative by Steinbauer (1989: 216) and Schrijver (1991: 217), but there is no nominal basis in Italic,
only Greek 'fodder'. LJV
2
tentatively reconstructs an iterative *gw orh;-eie- for Latin, whence *gworoje-,
which would have been analogically replaced by *gw ora} e-. The existence and productivity of *-aje- are unproven,
however. Vorare may rather be a deverbal d-present derived from an earlier subjunctive *gw erhrelo- or from the
PIE root aorist *g"'(e)rhr, cf. LIV
2
5.4 Compounds from a transitive base
anticipi'ire 'to occupy beforehand, anticipate', occupare 'to be the first, seize to oneself, take possession of' <
*kap-aje- to capio < *kap-i-. The meaning of these two compounds is atelic; the contrast of occupdre with
occipere 'to begin', which has a higher degree of transitivity, is evident. Anticipi'ire and occupiire have been
Latin Dcverbal Presents in -a-
325
regarded as deverbal formations by many scholars, but Steinbauer (1989: 139) returns to a dcnominal analysis,
based on a putative noun *op-kap- 'who takes on', whence *op-kap-a- 'to be an on-taker'> 'to occupy, attack'.
Since no compound *op-kap- is attested and since occupare belongs so closely to occipere, the deverbal analysis
seems more plausible.
educare 'to bring up, rear' < *-duk-aje- to ducere 'to lead' < *deuk-e/o-, educere 'to lead out'. The formal and
semantic aspects of the opposition resemble those of dicare versus dicere: zero-grade root in the a-present versus
full-grade root in the thematic present. Also, while educere usually refers to a single action of 'leading out',
educare typically stretches across a larger amount of time needed to 'rear' a person or 'tend' a plant.
profllgare 'to defeat decisively, crush, ruin utterly'. Found once in Plautus, Miles Gloriosus 230: conjidentiast nos
inimicos profligare posse, which is hard to analyse for its tclic or atelic semantics. Note, however, the derived
nounpr(jflzgator 'squanderer' (Tac.). The simplcxfllgere 'to strike down' is attested only a few times (Andronicus,
Accius), but also occurs in the compounds ajf/lgere 'to strike, cause destruction', confllgere 'to collide' and
ejj!Tgere 'to strike dead'. Thus, it seems quite possible that profllgare represents a deverbal present of atclic or
repetitive motion.
infriare 'to crumble ingredients in or on',.friare 'to pulverize, crumble',./i-icare 'to rub, chafe' to PIE *hhriH- 'to
cut'. The verbs .fricare andfriare are both clearly atclic and denote repetitive action. Traditionally, it is assumed
that the Latin forms arc based on three adjectives, *!lriH-o- 'cut' (whence .friare), *1/'riH-uo- (whence the
adjectivesfrivolus 'worthless, insignificant' and refdvus 'shredded') and *hhrilf-ko-. This is possible, but there are
alternatives. Adjectives in Pit. *-wo-, apart from colour adjectives and clear old cases such as saevus, arc mostly
deverbal to presents of the third and second conjugation (Leumann 1977: 303): assiduus, continuus, etc. With
transitive verbs, the adjective usually has a passive meaning. Thus, Proto-Latin *frlwo- 'small, shredded' might
represent a wo-adjcctive 'rubbed' to a present *jriH-e!o- which did not survive, but of which an atelic a-present is
found in.fri-a-re. The *-k- ofjricare (and its originally strong perfectfiicul,.frictus, later replaced by -avl, -atus) is
unexplained.
elegans picky, delicate' to Iegere 'to gather, collect' and eligere 'to select'. The retention of unstressed e
as an unreduced vowel suggest a more recent origin of elegans than of e/igere. Although only the present participle
is attested, it seems clear from its meaning that *e-legare denoted atelic 'to be selecting I selective', whence
'careful' and in Plautus 'decadent'. 'Selecting' or 'picking' is a often a repeated activity. A similar situation
obtains with this root in Celtic, where we find Middle Welsh dilein 'to destroy'< *di-leg-elo- but Old Irish legaid,
-/ega 'melts, melts away, perishes' < *leg-a-. It is not agreed that the Celtic root *leg- 'to melt' (plus Germanic
*lek- 'to leak) is the same as PIE *leg- 'to collect, read' (LIV
2
400 separates them). Still, since melting ice or
dripping liquid can be viewed as 'drops gathering', the connection is quite possible.
alligare 'to fasten one thing to another, put a noose round, hold together, impede the activity or movement of,
restrict', colligare 'to tie up, put in bounds, tie round', deligare 'to make fast by tying', obligiire 'to tie up, restrain
by tying; to assign for a specific purpose, pledge; to place under a moral or legal obligation, pledge oneself,
prae/igare 'to tie round; to fasten on to the front or extremity', re/igare 'to make fast, hold firmly in place',
subligare 'to fasten, gird up'. All have prevailingly atelic semantics referring to the repeated movement involved
with tying. The fact that the simplex ligare does not occur before Catullus suggests that it was metanalyzed from
the compound verbs. There is no certain cognate material in Italic, but in view of Albanian lidh 'ties' we may posit
a PTE root *li{!,-.
appellare 'to speak to, call on, invoke, appeal, designate', compe/lare 'to address, call upon, rebuke', interpel/are
'to intem.tpt, obstruct' to pe/lere 'to beat against, push, strike' <PIE *pel-n-hrelo- 'to approach'. The compounds
in -pel/are show atelic and repetitive semantics. They may be contrasted with appellere 'to drive to, touch' and
compellere 'to drive together, force' which are semantically closer to the synchronic meaning of pel/ere, and may
therefore be more recent compounds. Since -pellare mostly means 'to speak to', one may suspect that these
compounds were formed with the PIE meaning 'approach' of the root *pelhr, before pel/ere came to mean 'beat
against' (de Vaan 2008: 456).
applicare 'to bring in contact with, apply', complicare 'to fold together', displicare 'to scatter', explicare 'to
unfold, disentangle, extend', implicare 'to entwine, enfold, involve', replicare 'to fold back, bend back', all to
plectere 'to plait'. A simplex plicare is attested from Lucretius onwards; in view of Plautinc plicatrzx 'who folds
clothes', the verb must have existed in Old Latin, but the reduced vowel of plic- betrays it as a decompositional
formation. A denominal origin is usually assumed for supplicare 'to entreat, worship' to supplex 'suppliant'; but
the other verbs in -plicare have no nominal compound beside them, and it seems more likely that they are deverbal
presents of atclic or repetitive movement: 'folding' typically consists of several repeated actions.
326 Michie! de Vaan
assenti.JrT 'to agree, give confirmation; to flatter by agreeing' to sentTre 'to sense, feel'. The meaning of assentor is
atelic: all seven Plautine attestations have the meaning 'to flatter, agree on everything', and they take objects in the
dative or neuter pronouns. The compound assentTreli 'to agree, approve' means 'to agree' with something just said,
in a more conclusive way; hence it can be seen as telic. The difference between the two verbs was already illustrated
by Mcillct (1897: 85): Plautus, Amphitruo 702: Etiom tu quoque adsentaris huic? 'Arc you, too, going to back her
as well?' versus Plautus, Amphitruo 824: Mihi quoque adsunt testes, qui illud quod ego dicam adsentiant. 'I have
witnesses as well, who can confirm what I say'.
Since the active present participle of sen tire is sentient-, Leumann (1977) explains sen/entia 'opinion, purpose,
idea, sense' from a dissimilation of the first *i in expected *sentientia. Alternatively, it seems to me, sent-entia
could also be a remnant of a PIE root present or aorist of *snt- 'to notice'. In that case, Lat. sen/Ire would continue
*s(e)nt:ie-, whereas sent-ent-ia and ad-sent-arT would be built on a root formation *s(e)nt-.
a(b)sperni.JrT 'to refuse or reject, feel aversion for' to spernere 'to push aside, move away from, reject, scorn' <
PIE *slh!er-n-h
21
relo- 'to tread' . There is a clear pragmatic distinction in Plautus between asperni.JrT, which is
always used in a contrastive context (one person is being admitted, another one refused), and ,1pernere, which
means either 'to separate from' or 'to reject, scorn'. The meaning 'separate' is closer to the PIE meaning 'to tread',
and hence the main difference in meaning with asperni.Jrllies in the intensity of the act: while spernere is a single
act of pushing away or rejecting, a.spernclrl must be interpreted as 'continuously rejecting'.
conspici.Jrl 'to perceive, become aware oJ; get into one's sight', despicclrl 'to despise' and subspici.Jri- 'to form an
idea of; imagine, suspect', to specere 'to see, observe'. Each compound verb is attested beside a compound of the
third conjugation with the same preverb. There is a pragmatic difference between the members of each pair, as was
already illustrated by Meillet (1897: 85): conspicere, despicere and suspicere arc telic, whereas the compounds
in -spiciirT are atelic. In spite of the occurrence of adjectives such as despicus 'looking down' and prospicus
'looking ahead' in Naevius, it is of no avail to suggest that the compounds in -spicor are denominal, since this
would not explain the existence of both -spicio and -,spicor with the same prevcrbs, and with very similar meanings.
consterni.Jre 'to throw into confusion, confound; to drive or compel (by fear or other emotion)' to sternere 'to lay
out, spread' < PIE *ster-n--hrelo--. The compound verb clearly has atelic meaning, and the action of 'conJhsing'
also has a repetitive connotation. Contrast consternere 'to cover by strewing, pave; to bring down', the productive
compound of con plus sternere.
destini.Jre 'to arrange the purchase of, designate, settle on', obstini.Jre 'to set one's mind on', praestini.Jre 'to
bargain for, buy' to Proto-llalic *stVn-e/o- 'to stand' or transitive 'to put to stand' from PIE *st-n(e)-hr 'to stand,
arrange' (LIV
2
s.v. *stehr). The opposition between *stVnelo- 'stand' and -stini.Jre 'set' is reminiscent of that
between sedere 'to sit' and factitive sedi.Jre 'to make sit', where dynamic iJ adds a second role (an object) to an
otherwise stative verb. The putative present Pit. *stVn-e!o- is not attested but may have existed as a cognate of
other IE nasal presents to the root *sthr, in particular, Armenian stanam 'to acquire', Old Church Slavic stanQ 'I
stand, become', Gothic standan 'to stand', etc. The expected Proto-Latin vocalization from PIE *sf.n(e)-hrelo-
would be *stena/o-- (Schrijver 1991: 218-220), but since the root had the form *sti.J- or *sta- in most of its Pit.
forms, * sten- may have been changed to *stan- by analogy.
dissu/ipi.Jre 'to cause to separate in different directions, disperse, scatter', znsu/ipi.Jre 'to throw in' (Paulus ex Festa),
obsipi.ire 'to scatter, sprinkle', supi.Jre 'to pour, strew' (hapax Paulus ex Festo), to insipere, -io 'to throw in'. The
verb dissu/ipi.Jre can be regarded as atelic, while the remaining attestations of -supi.Jre from Paulus ex Festo are
used to translate iacio and have no context. We find obsipat 'sprinkles' once in Plautus. The action of 'sprinkling'
and 'scattering' is typically repetitive.
The root vowel of Latin is uncertain since supi.ire is only attested in glosses, and might be a backfonnation to
the other verbs in -supi.Jre. The best semantic match is Balta-Slavic *sup- 'to pour, strew, cover' (Lith. supti, OCS
suti).
5.5 Uncertain etymology and/or uncertain actionality
In the preceding sections, I have adduced what seem to me the clearest cases of devcrbal a-presents.
Once the possibility is accepted that an a-present without an obvious nominal model arose as a verb
with atelic actionality, it becomes possible to reconsider some additional presents. For some, the
etymology is unclear but the semantics of the verb render its appurtenance to this category conceivable.
In other cases, the a-present is only attested in glosses or in isolated quotations without context, which
exclude semantic proof of its atelicity. I will also reconsider a number of compounds with a nominal
first member. Far from claiming that all verbs discussed below are certain members of the atelic
Latin Devcrbal Presents in -a-
327
deverbal a-presents, their discussion is intended to explore the limits of the explanation proposed in
this paper.
5.5.1 Simplexes
cloare 'to clean' to cluere 'id.' Both stems arc only attested once: cloare in Scrvius (cloare 'inquinare'), cluere in
Pliny (cluere enim antiqui purgare dicebant). Their actionality therefore cannot be determined. The noun cloaca
'sewer' suggests that a verb cloare existed at an earlier stage, but a deverbal suffix *-ka is unusual. One might
regard cloaca as a feminine substantivization of an adjective *klowako- derived from an earlier stem *klowo-
'cleaning'. Of course, c/oare could be an immediate derivative of putative *k/owo-, as is actually assumed by Vine
(2006: 217f.).
ciinarf 'to make an effort, attempt'. These arc typical atclic semantics. The origin of long {j in the root is disputed.
Among the possible IE cognates are Middle Welsh di-gawn, di-chawn 'works; is able' and SCr. kaniti 'to intend',
which could go back to a root *kehrn- according to Matasovic 2009: 183. LJV
2
also connects Av. asanaoiti
'ascends', sanaj 'ascended', Gr. cy-Kovtco 'to be quick and active', 'servant', Arm. med. snaw 'was
raised', which would require a root variant PIE *k()nhr for c()narl.
fbdare 'to dig' (Paul. Fest.) to usual fi;dlrl 'to dig' and to .ft;dentes 'stabbing' (Ennius, Annales 504). 'Digging'
typically refers to a repetitive action. In view of the cognate verbs forms in Anatolian, Balto-Slavic and Tokharian,
the o-gradc of the root and maybe also the i-prescnt were inherited from PIE. For Hittite padda-/padd-', Melchert
(1984: 26) and Kloekhorst (2008: 654f.) assume that the root was *1/'e!od'hr. One might therefore be tempted to
explain Latinfodare as containing a reflex of the root-11nallaryngeal, in the wake ofRix 1999; yet in view of the
possibility thatfodare was made toftJdere at a more recent stage, like sonare to S(;nere, it seems safer to regard
.ft;dare as an atelic a-present.
manare 'to flow, pour, run, spread'. The etymology is unclear. Withinltalic, there is no noun from which manare
could have been derived, nor do I sec a good Indo-European connection. The root *mehr 'to give a sign, nod,
wink' does not fit semantically; a connection with madeii 'to be wet' is formally possible bul more complicated
(*mad(s)-n- ?), and semantically not evident, since 'to be wet' is not the same as 'to llow'.
rogare 'to ask' to regere 'to direct', PTE *h.)Yeg-. The verb is often considered to be denominal to a stem *rogo- or
*roga- 'direction'. The o-stem is attested in rogus 'funeral pyre' and also in ergii 'therefore, then', corgii 'forwards'<
*-rogo- 'direction' or 'stretching', butT see no immediate derivational path from 'direction' to 'asking'. Alternatively,
then, rogare continues an Italic a-present to an inherited Ogradc present with the meaning 'to turn to (someone)';
compare -pel/are 'to speak to' to *pelhr 'to approach'.
5.5.2 Preverbial compounds
indagare 'to track down, hunt; to search out, ascertain by investigation' to agere 'to drive'. The atelic meaning
renders a dcverbal origin as *endo-ag-iije- 'to drive onto, into' more likely than a denominal one to unattested
*endo-ag-o- 'hunt, drive'. The n-stcm indagii 'ring of huntsmen' (Verg.+) is attested too recently to warrant earlier
*endo-ago, and will rather be denominal to indagare itself. The lengthened grade of the root ag- also occurs in
amb-ages 'detour', co-agulum 'bond'.
occil/are 'to smash up', incflare 'to revile, abuse' to caedere 'to cut, hew, fell'. Steinbauer (1989: 196) has rightly
argued that occillare can hardly belong to occa 'harrow' but rather reflects (*)occllare < *ob-caelare 'to hew in
onto', which belongs to caedere - with intervocalic d to /. It is at first sight tempting to regard both verbs as
compounds of caelare 'to adorn, engrave', a denominative to caelwn 'chisel' < *kaidom. Yet the meaning of
occillare and inc flare has no direct connection with 'adorning, engraving', but is easier explained on the basis of
the meaning 'to cut, hew' of caedere. Hence, it is at least conceivable that we are looking at devcrbal a-presents to
caedere.
perfines 'pcrfringas' (Festus) looks like a subjunctive of a stem 'to break, fracture', which could then
be an atelic a-present to a putative the latter would be an exact formal match of Proto-Celtic *bina-
(Matasovic 2009: 65) <PIE *!h-n-fl-. The isolated attestation docs not allow us to judge the semantics.
confutare 'to keep fiom boiling over; restrain, diminish; abash, convict of error', rejutclre 'to check the activities of;
suppress; refute'. At least in the first two meanings, confiitare is atelic. There arc no nominal forms from
which jiitare could be derived, nor is a verbal origin immediately clear. The verb ejjiitlre 'to utter foolishly,
babble', which may be connected withfiitilis 'brittle, fragile, in vain', looks similar in form, but its meanings are
328 Michie! de Vaan
difficult to connect with Outside Italic, one might connect the Germanic root * baut- 'to beat', but its final
dental is different from that ofTtalic.
propagiire 'to reproduce, prolong, extend' to pangere 'to insert firmly, fix'. The verb is atelic, and its meaning
makes a deverbal interpretation as *pri5-piig-aje- easier than a denominal explanation from an unattested *pri5-pag-
o- 'offspring' or 'continuation'. The present was derived from the long vowel variant *pag- which is also attested
in compages 'framework, joint', pri5pages 'which continues', pri5piigi5 'offspring, space for planting', repiigula
[n.pl.] 'door-bars', pagus 'country district or community' and piigina 'column or page of writing'.
Jnsti'gare 'to incite, drive to action, urge on (in an action already being performed); to incite to anger, provoke; to
rouse (feelings)'. 'To urge on' clearly is atelic, while 'to incite to anger' has the factitive meaning of the type
sedare. Rather than a denominal verb to a putative *steig-(o-) 'sharp point', instlgclre could therefore be an atelic
present in *-aje- derived from verbal *steig-(e/o-). Verbal cognates are found in Skt. ati stigh- 'to overcome', a-
stig- 'to harm, penetrate, assail', tejate 'to sharpen' (cf. Lubotsky 2008), Greek cnil;w 'to sting, tattoo' and
Germanic *stikan- (OHG stehhan) 'to sting'. If the basic meaning of the root was stativc 'to be sharp', as is posited
by LJV
2
, then Latin *steig-a- can be seen as an atclic derivative which adds dynamicity and transitivity, as with the
type sedare.
praesti5lare!i 'to wait for, expect'. To PIE *stet- 'to arrange, place, set' (sec de Vaan 2008: 486)? If so, a
denominal origin is hardly likely: compare the meanings of Greek 'equipment', cn:oA.1] 'armour, dress',
from which praesti5lare cannot be derived. No other derivatives of the root *stet- appear in Latin, except possibly
stolidus 'insensible, stupid' and stoli5 'shoot (in plants)'.
amptruiire 'to execute a figure or movement', redamptruare 'to dance in response to the steps of a leader'. These
compounds have been explained as denominal to a putative *ambi-drew-o- 'running around', 'servant', but they
may equally well belong to a Proto-Italic verbal stem *drew-(e/o-) from PTE *dreu-; compare Skt. pr. dravati, 'to
run, hurry', drava- [adj.] 'running', YAv. drauuaiia- 'to take a run-up'. In that case, *ambi-drew-a- may be a
deverbal a-present of atelic movement.
5.5.3 Nominal compounds
I have excluded verbs with a nominal first member from my material, mainly because of the chance
that they were built to nominal compounds. Still, I think there is much to say for a deverbal origin of
tlsurpiire 'to carry out, make use of, take possession of, a clearly atelic verb, which is analyzed by
Vendryes (1910-1911: 300) and others as a deverbal present from an instrumental *t!sii plus the a-present
to rapere, -ii5.
1
Especially the occurrence of t/,nlcapere and tlstifacere renders the erstwhile existence of
*tisurapere plausible, more than a hypothetical nominal compound *usu-rap-o-. The same goes, mutatis
mutandis, for nuncupare 'to declare, appoint' < *nomo-kapaje-, which is generally seen as a denominal to
a compound *nomo-kap-.
A separate class is formed by the compound verbs in -igare and -lgare. With short internal i we find
iurigiire, itlrgare 'to quarrel', lltigare 'to litigate', navigiire 'to go by ship', piirigare, piirgiire 'to clean,
purify', which are derived from the consonant stems itls-, lit-, nav-, pur-. Word-internal long Tis found
in derivatives of i-stern nouns: castzgare 'to reprimand' ,fattgare 'to tire', vestzgare 'to follow the trail'
and investfgare 'to track down'. Both groups are usually explained as denominatives to putative
compounds in *-ex, *-igis < *-ag-s, *-agis (Weiss 2009: 402) or in thematic *-ago- (Dunkel 2000a),
and of course this cannot be excluded. Still, a deverbal origin is equally plausible. All of these verbs
represent atelic actions, so that we could be dealing with compounds in *-agare which were created
because telic agere was unsuitable.
l Watkins (1970: 323-328) has argued that the use of usurpiita and usurpare for the absence of a woman from
cohabitation with her man for three consecutive days in a year in order to avoid becoming his legal wife, must
rather continue *usu-rup- 'to break usus', from which the verb iJsurpare would be a back-formation. I agree
with Szemerenyi (1974: 181) who objects to the difficult if not impossible syntactic change presupposed by
Watkins' etymology. At the most, Szemcrcnyi admits, one might suppose that classical usurpare is due to a
coalescence of *usu-ruplire and usu-rapare.
Latin Devcrbal Presents in -ii-
329
6. Formal matters
6.1 Strong and weak perfect
The denominal a-presents have a weak perfect in -av-, -atus throughout in Latin; see the survey in
Mignot 1969: 269-279. Since denominal presents in my view did not have a01ists or perfects in PIE
(pace Meiser 2003: 134), *-aw- may have arisen at any time after *-w- became productive. This moment
must post-date the Latino-Sabellic unity, since Oscan applies a new tt-perfect to most denominal verbs,
whereas Umbrian has acquired a novel ns-perfect (Willi 2010).
Among the deverbal a-presents, the majority has weak perfect forms. Some have a strong perfect
in -u'i and/or -(i)tus: cubare, crepare, domare, iuvare, lavare, micare, necare (beside necavf), secare,
son are, tonare, vet are, .fi"icare. In Sabellic, a similar pair of forms is Umbrian prusekatu < *sek-a-
besides ppp. prusefYetu and asefYeta < *sek-eto-. Rix (1999) and Meiser (2003: 137-138) assume that
these a-presents continue PIE presents in *-hreie-, and that their perfects have replaced earlier aorists,
which therefore have to be reconstructed for PIE. Yet not all of these presents were inherited from PIE
-see the synchronic replacement of sonere by sonare --, nor did PIE form aorists to all roots of which a
present is attested. We are free to assume that some perfects in -uf, -itus were made to a Proto-Italic
present (gemere- gemuz, occulere- occuluf, etc.) and stayed in place while the present was joined or
ousted by a competing formation in *-aje-. If enough semantic specialization took place, a pair of
presents could be retained side by side, each with its own meaning, as in the case of lavare 'to have a
bath' and Iavere 'to wash'. Under this analysis, the kind of perfect which an a--present takes does not
allow any conclusions as to the relative age of that present in the prehistory of Latin.
6.2 Zero grade in the root
Ablaut differences between the derivative and its base verb are found with a number of verbs. Zero
grade instead of full grade occurs with dicare, educare, labare, and possibly also deliquare and mitat.
If these a-presents were deverbal, as I assume, their zero grade must be inherited from the base verb.
This would point to an ablauting root present or root aorist surviving at the time when the derivatives
arose.
6.3 Lengthened grade in the root
Three verbs show e-grade instead of full grade: celare, sedare, and venarL As argued above, sedare
and celari' might have been derived from the perfect, which for 'sit' is actually attested in sed'i. I am
less confident that venarf 'to chase' relies on a perfect *wen-<< *we-wn- 'has come to love"' likes',
but it is not impossible. An alternative origin for -e- is of course from a root aorist injunctive. Two
other verbs which have been compared are metarf 'to measure off' and legare 'to send as an envoy,
bequeath' (Isebaert 1992: 202f.). The former is only attested from Caesar onwards, and may well be
denominal to meta 'cone, limit'. The case of legare and the noun collega is less clear-cut: they could be
deverbal to Iegere 'to collect', but they could also be derivatives of lex 'law'.
Three other verbs display 6-grade. Whether c6narl" belongs here is uncertain, it may also contain
root-internal *-ehr. For sOlar/ and praesti5lare!f, one might look for a nominal origin of the long vowel,
see e.g. Isebaert 1992: 205. It is tempting to compare the well-known Greek iterative presents with 6-
grade in the root and a suffix *-a-, but the origin of co in this Greek category is still disputed. Many
scholars consider a nominal origin likely (e.g. Tremblay 1996), but Tucker ( 1990: 226--232) has
pointed out some internal Greek evidence in favour of a deverbal character of the iteratives in -aco. In
particular, all five Homeric instances of the type in -co-aco are very clearly deverbal: v c o ~ a c o 'to handle,
wield', ncoTUOf.Lat 'to fly', crrpco(paco 'to turn here and there, wander', rpconaco 'to turn, change' and
<pcoxaco 'to run, gallop'. Also, the motional semantics of these verbs recalls the atelic movement which
I understand to be basic to the Latin deverbals in *-aje-. This topic requires a separate investigation.
330 Michie! de Vaan
6.4 0-grade in the root
Quite a number of deverbal a-presents show short o in the root: dol are, domare,fodare,forare, rogare,
sonare, tonare, volare, vorare, and maybe Plautine vatare > Classical vetiire. As we have seen above,
the a-grade is not in all cases reliable. First of all, sanare can be excluded because its a can go back to
* swen-. The a offodclre is expected because its basisfodere < *fod-i- also has *o, as do some of the IE
cognates of this verb. For dolclre, valare, vorclre, and Plautine votare, it is impossible to determine
whether they contain *o or underwent rounding of *e. For volare and voriire, I regard *wel/r- as
somewhat more likely in view of the Latin tendency to unround *wo- to *wa- in open syllable
(Schrijver 1991: 460-465). This would leave doliire, damare, fodiire, fa rare, rogare, and tanare as
original o-grade verb stems. These, then, might be built on PIE a-grade presents with iterative,
intensive or resultative semantics, which alternated with zero-grade forms in the same paradigm; see
Klimmel2004 for the evidence, Kortlandt 2010: 373--382 for the semantics, and Kortlandt 2010: 383-
386 for an explanation of the ablaut a vs. zero.
7. Summary and conclusion
All verbs in sections 3 to 5 have atelic value, that is, the action has no natural endpoint. Section 3 deals
with the frequentatives in *-(e)s-a-, which involve repeated movement. In section 4, I discussed the
causatives to statives, which involve factitive verbs of movement. The following are the results for
sectionS.! to 5.4:
ATE_LIC:
POSITION: (-)cubare.
ACTIVITY:
PROLONGED: ambulare, amare, arare, celiire, -dinare, anticipare, occupare, (-)dicare, damare,
educare, iuvare, lahare, lavare, meare, mitat, imperare, properare, -pellare, assentar'i, (con)si5lar'i,
(re)sonare, aspernar'i, -spicar'i, ci5nsternare, -stinare, (can)tonare, venar'i, vetare, vo!are, varare.
REPETITIVE (event-internal plurality): crepclre, daliire, pri5fllgare, fa rare, (in)friiire, fricare, lzbare,
(-)micare, mulcare, -p!icare, runcare, secare.
FREQUENTATrVE, DISTRIBUTIVE (multiple-event plurality): calare, e!egans, dissipclre.
A TELl C) IELIC: -ligare, parare.
To be sure, the distinction between prolonged and repeated activity is sometimes an academic one. For
instance, pe!lare, arare, dicare, lavare, vorare may well have started out as repeated actions. The
presents -ligare and parare were originally atelic, but have adopted telic usages such as 'tie' and
'obtain' in due course. For the exceptional position of cubare (the only stative ), see section 5.1.
I conclude that it is possible to regard "atelic movement" as the central semantic component of the
deverbal suffix *-aje-. One might distinguish the following phases in the development of the function
of the suffix (though some functions will have overlapped, depending on the semantics of the verb):
I. Atelic movement (*-s-aje-, sedare)
2. Repetitive movement (jorare, secare)
3. Atelic activities (lavare, arare, crepare, parare)
4. Atelic derivatives from telic compounds (ci5nspicere --> ci5nspicarl)
5. Prefixed atelic verbs from a simplex (Iegere---> elegans)
Space restrictions prevent me from discussing the origin of the suffix *-aje- itself. The topic touches on
many disputed issues in the history of Latin (the a-subjunctive, the (b)a-imperfect, etc.) and Proto ..
Indo-European, and must be dealt with in a separate paper.
Latin Deverbal Presents in -ii-
331
References
Bammesberger, Alfred. I 996. Die maskulinen a-Stamme und der Verbaltyp occupare. In Alfred
Bammesberger and Friedrich Heberlein (eds.), Akten des VI!!. internationalen Kolloquiums zur
lateinischen Linguistik, 50--60. Heidelberg: Winter.
Barbclenct, Daniel. 1 913. De l 'aspect verbal en latin ancien et particulierement dans Terence. Paris:
Champion.
Beekes, Robert. 2010. Etymological Dictionary of Greek. Two volumes. Lcidcn/Boston: Brill.
Brugmann, Karl. 1916. Vergleichende Laut-, Stammbildungs- und Flexionslehre nebst Lehre vom
Gebrauch der Wortformen der indogermanischen Sprachen. Volume 2, Part 3/1. Strassburg:
Tri.ibner.
Cardona, George, Henry Hoenigswald, and Alfred Senn (eds.). 1971. Indo-European and Indo-
Europeans. Papers presented at the Third Indo-European Conference at the University of
Pennsylvania. Philadelphia: University of Pennsylvania Press.
Forssman, Bernhard. 2002. Etymologie im Thesaurus Linguae Latinae. Museum Helveticum 59.172-
187.
Garcia Castillero, Carlos. 2000. La formaci6n del lema de presente primario osco-umbro. Vitoria:
Universidad del Pais Vasco.
Haver ling, Gerd. 2000. On ,\co-verbs, prefixes and semantic functions: a study in the development of
prefixed and unprefixed verbs from Early to Late Latin. Goteborg: University of Goteborg.
~ ~ 2003. On prefixes and actionality in Classical and Late Latin. Acta Linguistica Hungarica
50.113-135.
van der Heyde, Klaas. 1926. Composita en verbaal aspect bij Plautus. Amsterdam University Ph.D.
dissertation.
---. 1932--34. L'aspect verbal en latin. Revue des Etudes Latines 10.326-336; REL 11.69--84; RilL
12.140-157.
Isebaert, Lambert. 1992. Spuren akrostatischer Prasensflexion im Lateinischen. In Oswald Panagl and
Thomas Krisch (eds.), Latein und Indogermanisch. Akten des Kol/oquiums der lndogermanischen
Gesellschafi, Salzburg, 23.-26. September I986, 193-205. Innsbruck: Institut filr Sprachwissen-
schaft der Universitat Innsbruck.
Kloekhorst, Alwin. 2008. Etymological Dictionary of the Hittite Inherited Lexicon. Leiden/Boston:
Brill.
Kortlandt, Frederik. 2007. Italo-Celtic Origins and Prehistoric Development of the Irish Language.
Amsterdam/New York: Rodopi.
--.2010. Studies in Germanic, Indo-European and Indo-Uratic. Amsterdam/New York: Rodopi.
Kravar, Miroslav. 1968. Zur Frage des lateinischen Verbalaspekts. Ziva Antika 18.49-66.
Kulikov, Leonid. 1999. Split causativity. Remarks on correlations between transitivity, aspect, and
tense. In Werner Abraham and Leonid Kulikov (eds.), Tense-Aspect, Transitivity and Causativity:
essays in honour of Vladimir Nedjalkov, 21-42. Amsterdam/Philadelphia: Benjamins.
--. 2008. The Vedic type patayati revisited: semantic oppositions, paradigmatic relationships and
historical connections. In Alexander Lubotsky, Jos Schaeken and Jeroen Wiedenhof (eds.),
Evidence and Counter-Evidence. Festschrift Kortlandt, vol. I, 323-342. Amsterdam/New York:
Rodopi.
Ki.immel, Martin. 2004. Zur o-Stufe im idg. Verbalsystem. In James Clackson and Birgit Anette Olsen
(eds.), Indo-European Word Formation, 139-158. Copenhagen: Museum Tusculanum.
Kurylowiez, Jerzy. 1956. L' apophonie en indo-europeen. Wroelaw: Zaklad im. Ossolinskich.
Leumann, Manu, 1977. Lateinische Laut- und Formenlehre. Miinchen: Beck.
LIV
2
= Rix, Helmut, eta!. 200 I. Lexikon der lndogermanischen Verben. Die Wurzeln und ihre Primiir-
stammbildungen. Zweite, erweiterte und verbesserte Auflage. Wiesbaden: Reichert.
Lubotsky, Alexander. 2008. The Indo-Iranian root *stig-. In Leonid Kulikov and Maxim Rusanov
(eds.), Indologica. T Ya. Elizarenkova Memorial Volume, Book I, 305-313. Moseow: RGGU.
Meillet, Antoine. 1897. L'expression de l'aoriste en latin. Revue de Philologie 21.81-90.
332 Michie! de Vaan
Meiser, Gerhard. 1986. Lautgeschichte der urnbrischen Sprache. lnnsbruck: Institut fi.ir Sprachwissen-
schaft dcr Universitiit Innsbruck.
1998. Historische Laut- und Forrnenlehre der lateinischen Sprache. Darmstadt: Wissenschaft-
lichc Buchgescllschaft.
2003. Veni Vidi Vici. Die Vorgeschichte des lateinischen Perfekt.'>ysterns. Mi.inchen: Beck.
Melchert, H. Craig. 1984. Studies in Hittite Historical Phonology. Gottingen: Vandcnhoeck and Ru-
precht.
Mignot, Xavier. 1969. Les verbes denorninatifs latins. Paris: Klincksieck.
Nussbaum, Alan, 2007. Latin present stems in -sa-: A possibly not so minor type. Handout, Kyoto,
September 2007.
Pedersen, Holger, 1921. Les formes sigrnatiques du verbe latin et le problerne dufutur indo-europeen.
(Kg!. Danske Vidcnskabernes Selskab. Historisk-filologiskc Meddelclser. Bind 3, nr. 5) Copen-
hagen : Host and Son.
Pinkstcr, Harm. 1990. Latin Syntax and Semantics. London: Routledge.
Praust, Karl. 2000. Studien zum indogermanischen Verbum. Ph.D. dissertation, Universitiit Munster.
Reinhold, Heinz. 1956. Zum lateinischen Verbalaspckt. ZVS 74.1-44.
Rix, Helmut. 1995. Einige lateinische Priisensstammbildungen zu Set-Wurzeln. In Wojciech Smoczyn-
ski (ed.), Kurylowicz Memorial Volume, Part one, 399-408. Cracow: Universitas.
---. 1999. Schwach charakterisierte lateinische Priisensstiimme zu Set-Wurzeln mit Vollstufe I. In
Heiner Eichner and Hans Christian Luschi.itzky (eds.), Compositiones Indogermanicae in memo-
riam Jochem Schindler, 515-535. Praha: enigma.
Romagno, Domenica. 2003. Azionalita e transitivita: II caso dei preverbi Iatini. Archivio glottologico
italiano 170.
Schrijver, Peter, 1991. The Reflexes of the Proto-Indo-European Laryngeals in Latin. Amsterdam/
Atlanta: Rodopi.
--. 2003. Athematic i-presents: the Italic and Celtic evidence. lncontri Linguistici 26.59--86.
--. 2006. Review of Meiser 2003. Kratylos
Sihler, Andrew. 1995. New Comparative Grammar of Latin and Greek. New York/Oxford: Oxford
University Press.
Sjoestedt, Marie-Louise. 1925a. Les iteratifs latins en -tare (-sare). Bulletin de !a Societe de
Linguistique de Paris 25.153--173.
----. 1925b. Les iteratifs latins en -tare (-sare) (Suite). Bulletin de Ia Societe de Linguistique de Paris
26.113-143.
Steinbauer, Dieter. 1989. Etymologische Untersuchungen zu den bei Plautus belegten Verben der
lateinischen ersten Konjugation. Unter besonderer Beriicksichtigung der Denominative.
Universitiit Regensburg Ph.D. dissertation.
Szemerenyi, Oswald. 1974. Review of Cardona, Hoenigswald, and Senn 1971. Journal of Linguistics
10.178-185.
Tucker, Elizabeth F. 1990. The Creation of Morphological Regularity: Early Greek Verbs in -eo, -ao, -66, -uo
and -fa. Gottingen: Vandenhoeck and Ruprecht.
de Vaan, Michie!. 2008. Etymological Dictionary of Latin and the Other Italic Languages. Leiden:
Brill.
--.Forthcoming. PIE i-presents, s-presents, and their reflexes in Latin. Glotta 87 [2011].
Vendryes, Joseph. 1910-1911. Sur quelques presents en -a- du verbe italo-celtique. Memoires de la
Societe de Linguistique de Paris 16.300-305.
Watkins, Calvert. 1970. Studies in Indo-European legal language, institutions, and mythology. In
Cardona, Hoenigswald, and Senn 1971,321--354.
Willi, Andreas. 2010. The Umbrian perfect in -n9-l-ns-. TPS I 14.
Wissmann, Wilhelm. 1932. Nomina postverbalia in den altgermanischen Sprachen: nebst einer
Vorunterschung iiber deverbative o- Verba. Gottingen: Vandenhoeck & Ruprecht.