LINGUIST List 15.264

Sat Jan 24 2004

Review: Historical Linguistics: Jasanoff (2003)

Editor for this issue: Naomi Ogasawara <>

What follows is a review or discussion note contributed to our Book Discussion Forum. We expect discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for review." Then contact Sheila Dooley Collberg at


  1. Mike Morgan, Hittite and the Indo-European Verb

Message 1: Hittite and the Indo-European Verb

Date: Sat, 24 Jan 2004 14:48:44 -0500 (EST)
From: Mike Morgan <>
Subject: Hittite and the Indo-European Verb

Jasanoff, Jay H. (2003) Hittite and the Indo-European Verb, Oxford 
University Press.

Announced at

Michael W. Morgan, Kobe City University of Foreign Studies

Key to typographical symbols used in this review:
* = reconstructed form
< derives from, goes back to
a_, e_, o_ = a, e, o with macron
a^, e^, u^ = a, e, u with breve
u'' = u with umlaut
a', e', i', o' = a, e, i, o with acute accent
o> = o with right (Polish) hook
i#, u# = i, u with subscript arch
c^, s^ = c with wedge, s with wedge
r^ = r with under-ring (i.e. syllabic r)
h^ = h with subscript breve
h1, h2, h3 = h plus subscript 1, 2, 3


Modern discussion of Indo-European (IE) linguistics, be it phonology,
morphology or even syntax, must take the data provided by Hittite into
account. While too much discussion of IE simply appends the data of
Hittite to the traditional framework obtained on the basis of the
'inner IE' languages (i.e. the state of the art before Hittite and
Tocharian were discovered and analyzed), making only relatively minor
modifications to this framework, the monograph under discussion is a
thorough rethinking of the IE verbal system, taking the data of
Hittite as a starting point.

The current monograph is clearly intended for specialists in IE verbal
morphology, but could potentially be used by (very) advanced students
of IE Studies (someone with a thorough grasp of the sections on verbal
morphology provided by the modern standard handbooks, such as Beekes
(1995), Lehmann (1993) and Szemere'nyi (1987), for example),
especially if they had a reasonable background in Hittite as well.


Chapter One examines the problem of Hittite h^i- conjugation, and its
relation to IE verbal conjugation. Hittite has two conjugation
classes: the -mi class and the -h^i class. While the former
corresponds directly to the Proto-Indo-European (PIE) primary active
endings, the latter class lacks close counterparts in the other IE
languages. From a purely formal point of view Hittite possesses the
following active system:

Present endings:
 First person singular: -(h^)h^i < PIE *-h2e+i
 Second person singular: -(t)ti < PIE *-th2e+i (with analogization)
 Third person singular: -i < PIE *-e+i or *-o+i
 First person plural: -wen(i) < -mi conjugation
 Second person plural: -(t)ten(i) < -mi conjugation
 Third person plural: -anzi < -mi conjugation

Preterite endings:
 First person singular: -(h^)h^un pre-Hittite *-(h^)h^a
 + mi-conjugation ending -un
 Second person singular: -(t)ta < PIE *-th2e
 Third person singular: -s^
 First person plural: as above
 Second person plural: as above
 Third Person plural: -er < PIE *-e_r

Theories of the origin of the -h^i conjugation can be divided into the
following groups: (1) the perfect theory (the majority view), which
derives the h^i-conjugation from classical PIE perfect (PIE present
singular endings *-h2e, *-th2e and *-e, as well as third plural ending
*-e_r in the above derivations are all perfect endings), (2) the
middle theory, which derives the h^i-conjugation from the PIE middle
(by meaning, verbs with h^i-conjugation in Hittite often correspond to
verbs that inflect as middles elsewhere in IE; also present
conjugation third person singular could derive from PIE middle *-o+i
rather than perfect *-e+i), (3)the Indo-Hittite theory, which sees the
h^i- conjugation as a purely Hittite innovation outside of IE, (4) a
mixed theory, which derives the h^i-conjugation from a 'sister' of
both the perfect and the middle, all of which go back to an early IE
proto- middle, and (5) the thematic conjugation theory, which derives
the first person singular *-o_ from PIE *-o-h2 and the third person
singular from PIE perfect *-e. Jasanoff argues against each of these
theories, and that a large number of Hittite h^i-conjugation verbs
''correspond in root etymology and stem structure to ORDINARY ACTIVE
PRESENTS OF A VARIETY OF FORMAL TYPES'' (p. 28; emphasis in original).

Chapter Two provides morphological preliminaries to the discussion
provided in subsequent chapters. In this preliminary, Jasanoff
discusses the perfect and middle and their relation to the h2e-
conjugation (Hittite -(h^)h^i). Jasanoff argues that the ancestor of
the perfect, at some time in the prehistory of the parent language,
was neutral as to tense, conveying both present and past stative
meanings. Later in the parent language the two meanings were formally
distinguished, by, for example, addition of secondary active ending in
the third singular and accent shift and root apophony in the third
plural. The perfect in the prehistory of IE is linked both formally
(derivationally) and semantically with the middle. This link is
attributed by Jasanoff to a common ancestor in the prehistory of PIE,
where the third persons differed (*-e in the perfect and *-o in the
middle) and the other forms were largely identical. This
'proto-middle' was ''at least broadly comparable in function to the
middle endings'' of later IE languages (p. 59). In the subsequent
history of PIE it is almost entirely the forms of the middle that
undergo innovation, while those of the perfect remain largely
unchanged. Jasanoff closes this chapter with a discussion of the
thematic first singular form *-o-h2, a form closely linked with the
prehistory of the perfect and middle.

The complete set of PIE forms involved can be reconstructed as follows:
Sg. 1 *-h2e
2 *-th2e
3 *-e
Pl. 1 *-meH (?)
 2 *-(H)e (?)
 3 *-e_r (< **-ers), *-r^(s)

Secondary Primary
 Sg. 1 *-h2e *-h2e-r
 2 *-th2e *-th2e-r
 3 *-o, *-to *-o-r, *-to-r
 Pl. 1 *-medhh2 *-medhh2(-r?)
 2 *-dh(u)u#e *-dh(u)u#e(-r?)
 3 *-e_ro, *-ro, *-nto *-(e_)ro(-r?), *-nto-r

In Chapter Three Jasanoff discusses the link between the h2e-
conjugation and root presents in PIE. 'Molo_- presents' (< PIE *melh2-
'grind') form a group with etymological links to Hittite h^i-
conjugation presents (malla/i-, malliya/i-). Verbs of this group often
have an 'expressive' character, and ''the majority fall into two
semantic groups: verbs of motion ..., and verbs of vigorous or violent
activity'' (p. 76). This group, Jasanoff argues, was probably
considerable in size in PIE. Also among root presents are the class
containing Hittite s^a_kk-/s^a^kk- /s^ekk- 'know' which is linked to
the PIE Narten present *se_kH-/*sekH-. Verbs in this group are notable
for their shared apophonic behavior, a behavior not shared by molo_-
presents. Jasanoff closes the chapter with a discussion of how PIE
differentiated the present proper from the imperfect/injunctive.

In Chapter Four Jasanoff discusses the link between the
h2e-conjugation and i-presents. Molo_-presents were not the only
active presents that inflected with the h2e-series of endings; another
group is the i- presents. A standard example is Hittite da_i/tiyanzi
'put', which has a strong stem in -ai- and a weak stem in -i- or
-iya-. Like the molo_- presents, this is a sizeable group of verbs;
the vast majority whose etymology is known go to PIE roots of the form
*(C)CeH- (where H is most commonly *h1). The -i of the third person
singular in Hittite seems to go back to a PIE *-i- or *-i#-; again
most of those whose etymology is known go to an i#e/o-present
somewhere outside of Anatolian, ''precisely the sequence that would
have resulted from the thematization of a stem-final element *-i-''
(p. 97). Jasanoff's conclusion is that the ancestors of this group in
the parent language were athematic i-presents with h2e- conjugation
inflection. Another type of i-inflection is represented by Hittite
mema/i- 'say'. This type has: (1) a strong stem in -a-, usually with
shortened -ai-, (2) a weak stem in -i-, and (3) a preference for
sigmatic endings, including second and third singular preterite
-is^/-es^ or -is^ta/-es^ta. The final, and by far most numerous group
of i-presents is the 'duratives' in -anni/a (e.g. iyannai/iyanniyanzi
'start marching, get underway'). This group would seem to derive from
a group with an *-nh2-i (and corresponding inner-IE *-nh2-i#e'/o')
suffix. The reason for derived h2e-conjugation i-presents may lie in
an iterative or durative nuance given to the stem by the h2e-
conjugation endings.

In Chapter Five Jasanoff traces the connection between the h2e-
conjugation and other characterized presents. One such group of
characterized presents is the reduplicated class represented by
Hittite mimma- 'refuse', which can be set up with an athematic
h2e-conjugation paradigm. Closely related is the class of iteratives
in -s^s^(a)-, a closed class represented by such verbs as h^alzis^s^a-
'call'. The last major group of h^i-conjugation verbs with clear
etymological ties to characterized presents in the parent language are
denominative factitive type, e.g. Hittite newah^h^- 'make
new'. Jasanoff also sees u-presents, such as PIE *dhe'nh2-u- 'move
off' as likely candidates for h^i- conjugation type. These presents
share with i-presents the tendency to become thematic in the later

In Chapter Six and Seven Jasanoff examines aorists of the h2e-
conjugation. If PIE h2e-conjugation presents (and imperfects) are
'denatured protomiddles', then we should expect to find similar h2e-
conjugation aorists, which present themselves in Hittite as ordinary
h^i-verbs. That is to say, they are simply h^i-conjugation root
presents in Hittite, but with etymological links to root aorists
elsewhere in the IE family. Outside Anatolian, the root aorists in
question should be identifiable based on their *o:*e ablaut (and, less
likely, by the h2e-conjugation reflexes). The first such type found is
the 'stative- intransitive' type, as for example PIE *logh-/*legh-
'lay down', Hittite lak- 'knock out, bend'. The second h2e-
conjugation aorist type is the 'presigmatic' aorist, exemplified by
nai- 'turn, direct'. This group was a mixed category, with the
non-third singular endings being non- sigmatic and taking the perfect
endings. The creation of a fully sigmatic active indicative is a
common innovation of the inner-IE (i.e. non-Anatolian, non-Tocharian)
languages, which thus form a proper subgroup of IE as a whole.

Chapter Eight is a retrospective of the argumentation provided in the
monograph. Jasanoff argues that the post- h2e-conjugation model of the
verb is extremely conservative: three persons, three numbers, active
and mediopassive, fully grammaticalized contrast of present/imperfect
and aorist, a perfect (with both active and middle, and four moods
(indicative, imperative, optative and subjunctive). The new system is
novel on the formal level; grammatical actives could have either the
normal active mi-conjugation endings, or the perfect h2e- conjugation
endings. Which it had was determined by its IE-internal history; the
distinction was functionally opaque (like the distinction in English
between strong and weak verbs), and thus unstable in the later
languages. Jasanoff closes with a list of new problems raised by the
h2e-conjugation theory.

Appendix 1

This appendix provides a prehistory of the thematic conjugation. We
have already seen the h2e-conjugation connection with the first
singular o-h2 ending. The entire thematic paradigm for the presents
are exemplified by bhe'r- e-h2e (> o-h2), bhe'r-e-si, bhe'r-e-ti,
etc. For thematic presents correlated with presigmatic aorists, such
as *ne'iH- e/o- 'lead', Jasanoff opts for a subjunctive origin.

Appendix B

This appendix treats the IE perfect u#o'id-e. In particular, the lack
of reduplication in any of the IE language reflexes is accounted for
as an inner-IE neologism, a back-formation from its own middle.


Jasanoff's monograph is a very welcome addition to IE studies.
Argumentation is both thorough and convincing. A wealth of forms from
the full range of IE languages is provided in course of the
argumentation. The approach is both comparative and empirical; through
Jasanoff's arguments the forms speak for themselves, and we see the
descriptive situation of the parent language. While too difficult a
text to be used in an introductory course on the IE verbal system, as
stated at the outset, it might recommend itself to the more advanced
student (and certainly to the IE scholar).

One quibble over terminology. In Chapter One, Jasanoff presents
numerous reasons for not arguing against a perfect-based origin for
the h2e-conjugation. In part this is due to the lack of reduplication
in the Hittite h^i- conjugation, a reduplication which is so much a
part of the perfect in inner-IE languages as Greek or Sanskrit. Also
Jasanoff argues that a perfect origin would mean that the h^i-
conjugation had spread to an apparently arbitrary set of present
stems. While arguing against positing a meaning function for the form
of the IE -h2e conjugation, Jasanoff then goes on to call it a
'protomiddle' or Urmedium (p. 146), thus taking sides in the perfect
versus middle argument. He argues that, due perhaps to its special
functional position, the perfect parted company with the main line of
development of the h2e-conjugation. As opposed to the protomiddle, the
perfect is seen as neoactive.

We, however, would rather see the h2e-conjugation as a 'protoperfect'.
It is, after all, the middle which is formally remodeled; the perfect
remains largely unmodified. Additionally, although generalized in
Greek and Indo- Iranian, reduplication is spotty at best in the other
inner- IE languages. On this basis we could equally well assume a PIE
reduplication with reduced range (something like the case of Latin
with a relatively few reduplicated forms), which subsequently became
generalized in some groups, and lost in others. Hittite would be one
of the latter groups (with such as wewakk- 'demand' remaining as rare
holdovers of the original PIE situation). As Perel'muter (1977) shows,
most perfects of Homeric Greek are without corresponding presents (or
else with derived, rather than root, presents); many have present
meaning. There seems to be an almost complementary distribution
between perfect and present stems; a given stem has either one or the
other set of endings. Such a situation would be exactly like what we
find in Hittite.

The most notable shortcoming of the book is its absence of a topical
index, or of an index of authors cited. It does, however, possess an
index of forms cited. This shortcoming makes it extremely doubtful
that the text could be managed by anyone without a comfortable
knowledge of most of the pertinent literature, and of the terminology
which has grown up around the IE verbal system which Jasanoff uses
repeatedly (e.g. Narten present, tuda'ti presents, etc.)

Although it was impossible and impractical to check all the forms
cited by Jasanoff, one mistake was noted in passing. In the Old Church
Slavonic reflex of IE u#edh- given on page 225, the form of the
present is given as vezo> rather than vedo>. (The correct form is,
however, given on page 174.)

This monograph is clearly must reading for anyone seriously interested
in PIE verbal morphology, and is doubtlessly destined to become a
standard reference in the field for years to come.


Beekes, R. S. P. (1995). Comparative Indo-European Linguistics: An
Introduction. John Benjamins.

Lehmann, W. P. (1993). Theoretical Bases of Indo-European Linguistics.

Perel'muter (1977). Obs^c^eindoevropejskyj I grec^eskij glagol. Nauka.

Szemere'nyi, O. (1987). Introduccio'n a la Lingu''i'stica Comparativa.
Editorial Gredos. [Spanish translation of Szemere'nyi (1970).
Einfu''hrung in die vergleichende
Sprachwissenschaft. Wissenschaftliche Buchgesellschaft.]


Michael W. Morgan has a doctorate in Slavic Linguistics, and currently
teaches at Kobe City University of Foreign Studies, Kobe, Japan. An
Indo-Europeanist by training, he has been involved in sign language
research for the past ten years, but also continues to conduct
research on Indo- European languages as well.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue