Eitan Grossman

No case before the verb, obligatory case

after the verb in Coptic*
Abstract: This paper presents a hitherto unnoticed fact about the coding of gram-
matical relations in Coptic: while postverbal core arguments must be overtly case-
marked (or “flagged”), preverbal core arguments are never case-marked. This
feature extends the “no case before the verb in northeastern Africa” generaliza-
tion (König 2008; 2009) to the northeastern Mediterranean. Moreover, the analy-
sis presented here reveals Coptic to be another case of an uncommon system of
core argument marking, namely, “marked S/A vs. marked P”.

1 No case before the verb in Coptic

Despite the fact that Coptic is a well described and abundantly attested language,
with a dozen or so dialects (Funk 1988), a simple yet important feature of its
grammatical structure has gone unremarked in grammatical descriptions: there
are no core case distinctions before the verb, and postverbal core arguments must
be case-marked.1 This feature extends König’s (2008; 2009) “no case before the
verb in northeastern Africa” generalization geographically to the northeast Medi-
terranean. It also extends it genealogically, adding Egyptian-Coptic to Berber as
Afroasiatic languages with this feature.2 Finally, Coptic corroborates König’s gen-
eralization, according to which “if there is no case distinctions before the verb,
then preverbal participants occur in the morphologically unmarked form” [i.e.,

*  This paper grew out of joint work by the author and Giorgio Iemmolo (Zurich), funded in part
by the Swiss National Science Foundation (Grant IZK0Z1_146024). I would like to thank Denis
Creissels, Zygmunt Frajzyngier, Tom Givón, Tom Güldemann, Martin Haspelmath, Christa König,
Sebastian Richter, Ewa Zakrewska, Yael Ziv and other participants in the Association for Linguis-
tic Typology 10th Biennial Conference (Leipzig 2013) for their helpful comments and criticism.
1  I assume here a broad definition of the comparative concept “case marker”, roughly corre-
sponding to the notion “flag”, “relator”, or “dependent-marker”. Such definitions are common
in typological studies of case (e.g., Haspelmath 2008; Siewierska & Bakker 2008), and it is in this
sense that the term “case” is used in this paper.
2  Whether Berber “states” ought to be described in terms of case is disputed; see Mettouchi &
Frajzyngier (2013) on Kabyle, as well as Arkadiev’s reaction (fc).

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
204   Eitan Grossman

the morphosyntactically simplest form, EG] (2008: 281), but extends it to lan-
guages which do not have ergative or marked-nominative coding.
The structure of this paper is as follows: in § 2, I present a bare-bones account
of the encoding of grammatical relations in Coptic3 and in §  3 some frequency
data. In § 4, I argue that Coptic does not have a marked nominative system. In
§ 5, I provide a brief note on diachrony, and in § 6, König’s “no case before the
verb” generalization in African languages is briefly sketched. All examples,
unless noted otherwise, are taken from the earliest documentation of Bohairic,
the northernmost Coptic dialect (Grossman 2009a).

2 Encoding grammatical relations in Coptic:

indexing and case-marking
Grammatical relations are language-specific categories, often (but not exclu-
sively) encoded by means such as indexation (“agreement”), flagging (“case”),
and linear order (Frajzyngier & Shay 2003; Payne 2013). Behavioral properties,
e.g., control of reflexivization, will not be dealt with here.
Grammatical relations in Coptic are encoded by means of four main strate-
gies: indexation, argument incorporation, case marking, and linear order. In this
section, I refer only to main (i.e., non-subordinate) intransitive and monotransi-
tive4 verbal clauses of a particular construction type, the so-called “non-durative
pattern” or “tripartite conjugation” (Polotsky 1960). This morphosyntactic5 con-
struction type comprises up to four “slots”:

3  A more detailed account of Coptic grammatical relations, in the context of Differential Subject
Marking, is given in Grossman & Iemmolo (2014+). The present article provides only enough
background in order to substantiate the claim about case-marking and linear order.
4  This is intended to exclude ditransitives, on the one hand, and bivalent intransitives, on the
other. Transitive clauses are defined according to Haspelmath (2011b) and Lazard (2002): broad-
ly, a monotransitive clause is a bivalent clause that has A and P as its core arguments; intransi-
tives can be either monovalent (one argument) or bivalent, if the arguments are not A and P. In
turn, A and P are defined, respectively, as “the argument of the major two-argument construction
that represents the agent when the construction expresses an action” and “the argument of the
major two-argument construction that represents the patient when the construction expresses
an action” (Haspelmath 2011b).
5  I use the term “morphosyntactic” instead of “morphological”, since the latter assumes that
the notion “word” is well defined, either as a cross-linguistic comparative concept or as a de-
scriptive category of Coptic (Haspelmath 2011a). For the description of Coptic, the term “bound
group” (Layton 2004) is more than adequate.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   205

Table 1: The constructional scheme of a Coptic verb

TAM auxiliary A/S expression lexical verb P expression

a f t amio
h
f

pst 3sgm create 3sgm

‘He created him.’

This is a fairly simplified representation of the structure of the verb, since the
TAM auxiliaries can be discontinuous (e.g., the future marker e…e), lexical argu-
ments can occur in the A/S and P slots, and both preverbal and postverbal lexical
arguments are very frequent in discourse. However, it is important to keep this
basic structure in mind, since it is to this construction (rather than the lexical
verb) that I refer when I use the terms “verb (V)”, “preverbal” or “postverbal.”

2.1 Argument Indexing

Intransitive clauses have a maximum of one argument index (ex. 1); in monotran-
sitive clauses, one or two arguments can be indexed on the verb (ex. 2). There is
no implicational relationship between A and P indexing, since an A index can
occur without a P index, and vice versa (see 2.5 below).

(1) ⲁⲓⲭⲱⲡ
a-i-khôp S
1sg-hide
‘I hid’ (Early Bohairic, Genesis 3:10).

(2) ⲁϥⲑⲁⲙⲓⲟϥ
a-f-thamio-f A+P
pst-3sgm-create-3sgm
‘He1 created him2’ (Early Bohairic, Genesis 1:27).

In terms of alignment, indexing is mixed accusative-neutral, depending on a

complex set of phonological and morphosyntactic factors, which are irrelevant
to the present discussion. Example (2) shows neutral alignment (A=P) in index-
ing for this particular constellation of verbal construction, tense, and person.
Example (3) shows accusative alignment in indexing: the 1sg A index (-i), which
is the same as the 1sg S index (cf. ex 1 above), differs from the 1sg P index (-t).

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
206   Eitan Grossman

(3) ⲉⲓⲧⲱⲛⲧ
e-i-e-tôn-t S=A≠P (cf. ex. 2)
fut1-1sga-fut2-raise-1sgp
‘I will raise myself up’ (Early Bohairic, Nahum 3:5).

2.2 Argument Incorporation

Both lexical subjects (A/S) and objects (P) can be incorporated into the verb,6 but
it is rare for both lexical A and P to be incorporated into the same verb. Incorpo-
rated P arguments are bound to the lexical verb, which often shows a prosodically
reduced form. This prosodically reduced verb form cannot occur as a free mor-
pheme. For example, in examples (6–8), the forms of the lexical verbs without
incorporated P are, respectively čimi, khô, and thamio.
(4) S-incorporation
ⲁⲣⲟⲟⲩϩⲓ ϣⲱⲡⲓ
a-roouhi šôpi
pst-evening become
‘It became evening’ (Early Bohairic, Genesis 1:8).

(5) A-incorporation
ⲁⲫϯ ⲑⲁⲙⲓⲟ ⲛⲛⲓⲑⲏⲣⲓⲟⲛ
a-ph[nou]ti thamio n-ni-thêrion
pst-G[o]d create [Link]-beast
‘God created the beasts’ (Early Bohairic, Genesis 1:25).

6 The term “incorporation” is used somewhat unconventionally here, since the arguments
bound within the morphosyntactic verb can be referential noun phrases as well as “bare” noun
lexemes. In this respect, Coptic differs from constructions described as “incorporation” in other
languages, but is similar to the way that Boumaa Fijian has been described by Dixon (1988); see
also Aikhenvald (2007). On the other hand, Coptic unambiguously treats the verb and the object
as a single bound group, by use of a prosodically-reduced verb form that cannot occur on its own.
As such, the incorporation analysis will be retained here.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   207

(6) P-incorporation
ⲙⲡⲟⲩϫⲉⲙⲃⲟⲏⲑⲟⲥ ⲇⲉ ⲛⲛⲁⲇⲁⲙ
mp-ou-čem-boêt os=de
h
nn-adam
[Link]-3pl-find-helper=conn acc-Adam
‘A helper was not found for Adam’ (Early Bohairic, Genesis 2:20).

(7) A+P-incorporation
ⲁⲡϭⲥ ⲫϯ ⲭⲁⲟⲩⲙⲏⲓⲛⲓ ⲛⲕⲁⲓⲛ
a-pc[oi]s ph[nou]ti kha-ou-mêini n-kain
pst-the.L[or]d G[o]d [Link]-mark acc-Cain
‘The L[or]d G[o]d put a mark on Cain’ (Bohairic, Genesis 4:15).

(8) ⲁⲫϯ ⲑⲁⲙⲓⲉϯⲫⲉ

a-p [nou]t thamie-ti-phe
h i

pst-G(o)d create-def.m-heaven
‘God created the world’ (Early Bohairic, Genesis 1:1).

2.3 Case-marking

Coptic is not traditionally described in terms of case marking. Nonetheless, Coptic

has adpositions and other flags that code grammatical relations. As noted above,
I assume here a broad definition of the comparative concept “case marker”,
roughly corresponding to the notion “flag”, “relator”, or “dependent-marker”.
Non-incorporated postverbal lexical subjects (S/A) and objects (P) must be
overtly case-marked. The accusative marker (n-) is seen in example (9),7

(9) Accusative
ⲁϥϫⲓ ⲛⲟⲩⲕⲁϩⲓ ⲉⲃⲟⲗ
a-f-či n-ou-kahi ebol
pst-3sgm-take [Link]-earth out
‘He took earth’ (Early Bohairic, Genesis 2:7).

7  The accusative case marker was grammaticalized from a highly polyfunctional preposition
associated with locative, instrument, source, and a number of other functions. See Winand
(2014, in this volume) for a discussion of its sources.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
208   Eitan Grossman

The opposition between object incorporation and accusative case marking is gen-
erally seen as a case of Differential Object Marking motivated by the interaction of
referentiality and topicality rather than, e.g., animacy or definiteness (e.g., Eng-
sheden 2008; Grossman 2009b; Winand 2014, in this volume).
The nominative marker nče (nci in other dialects, e.g., Sahidic) is found in
example (10).

(10) Nominative
ⲁⲩⲭⲟⲡⲟⲩ ⲛϫⲉⲁⲇⲁⲙ ⲛⲉⲙⲧⲉϥⲥϩⲓⲙⲓ
a-u-khop-ou nče-adam nem-tef-shimi
pst-3pl-hide-3pl nom-Adam with-his.f-woman
‘Adam and his wife hid’ (Early Bohairic, Genesis 3:8).

Few if any Copticists have described nci/nče as a nominative marker. It has

typically been described as a preposition (Layton 2004), extrapositive subject
marker (Loprieno 2000), or focus marker (Reintges 2004). However, it does not
have other properties typically associated with prepositions in Coptic, and it is
not strongly associated with focus in most dialects. The descriptive label “extra-
positive” (or perhaps “postpositive”) subject marker would be adequate, but it is
perhaps better to have a more specific label. The term “nominative” seems appro-
priate enough, since nci/nče is highly grammaticalized for a particular function,
marking postverbal lexical A/S arguments, and has a high token frequency, which
indicates that it is not just an “afterthought marker”, along the lines of English
“namely” or Seneca neh.8
The co-occurrence of both lexical nom and lexical acc is possible but rela-
tively rare:

(11) ⲁϥⲑⲁⲙⲓⲟ ⲛϫⲉⲫϯ ⲙⲡⲓⲧⲁϫⲣⲟ

a-f-thamio nče-ph[nou]ti m-pi-tačro
pst-3sgm-create nom-G[o]d acc-def.s-firmament
‘nomGod created accthe firmament’ (Early Bohairic, Genesis 1:7).

This rarity stems from the nature of discourse: transitive clauses with more than
one lexical core argument tend to have low text frequency across languages
(DuBois 1987).

8  I was made aware of this marker, which looks like a candidate for antitopic marker status, by
Wallace Chafe in a lecture (Leipzig, 12/8/2013).

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   209

2.4 Preverbal lexical arguments and case-marking

Preverbal lexical arguments are not case-marked.

(12) Preverbal S
ⲁⲃⲉⲗ ⲇⲉ ⲁϥϣⲱⲡⲓ
Abel=de a-f-šôpi
Abel=conn pst-3sgm-become
‘Abel became…’ (Early Bohairic, Genesis 4:2).

(13) Preverbal A
ⲁⲇⲁⲙ ⲇⲉ ⲁϥϯⲣⲁⲛ ⲉⲛⲓⲧⲉⲃⲛⲱⲟⲩ
Adam=de a-f-ti-ran e-ni-tebnôou
Adam=conn pst-3sgm-give-name [Link]-animals
‘And Adam named the animals’ (Early Bohairic, Genesis 2:20).

The following example shows that the absence of case marking is not limited to
proper names.

(14) Preverbal P
ⲡⲓⲭⲁⲕⲓ ⲇⲉ ⲁϥϯⲣⲉⲛϥ ϫⲉⲡⲓⲉϫⲱⲣϩ
pi-k aki=de
h
a-f-t -ren-f
i
če-pi-ečôrh
def.m-darkness=conn pst-3sgm-give-name-3sgm quot-def.m-evening
‘The darkness, he named it “evening”’ (Early Bohairic, Genesis 1:5).

Preverbal lexical arguments are often marked topics (ex. 15), but they can occa-
sionally be focal as well (ex. 16). Interestingly, topical preverbal arguments tend
to be accompanied by the particle de, while focal preverbal arguments tend to be
directly followed by the verb.

(15) ⲁⲇⲁⲙ ⲇⲉ ⲁϥⲥⲟⲩⲉⲛⲧⲉϥⲥϩⲓⲙⲓ

Adam=de a-f-souen-tef-shimi
Adam=conn pst-3sgm-know-his-woman
‘As for Adam, he knew his wife’ (Early Bohairic, Genesis 4:1).

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
210   Eitan Grossman

(16) ⲛⲓⲙ ⲁϥⲧⲁⲙⲟⲕ

nim a-f-tamo-k
who pst-3sgm-inform-2sgm
‘Who told you?’ (Early Bohairic, Genesis 3:11).

2.5 Interim summary

Coptic has three9 main strategies for coding lexical A/S:

Table 2: Indexing and case for lexical A/S

Indexing Case
Incorporated A/S no no
Preverbal A/S yes no
Postverbal A/S yes yes

Coptic also has three main strategies for coding lexical P:

Table 3: Indexing and case for lexical P

Indexing Case
Incorporated P no no
Preverbal P yes no
Postverbal P no yes

Several generalizations can be made:

1. All postverbal core arguments must be case-marked.
2. Preverbal and incorporated core arguments are never case-marked.
3. All preverbal core arguments, as well as postverbal subjects, entail indexing.
4. There is no implicational relationship between A and P indexing; all combi-
nations of A and P indexes and lexical arguments are possible:

9  There are other, relatively infrequent, construction types, but they will be ignored here.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   211

Table 4: Co-occurrence of A/P indexes and lexical arguments

Indexed P Incorporated P
Indexed A a-f-t amio-f
h
a-f-thamie-th-pe
pst-3sgm-create-3sgm pst-3sgm-create-def.f-heaven
‘He created him.’ ‘He created heaven.’

Incorporated A a-ph[nou]ti thamio-f a-ph[nou]ti thamie-thpe

pst-G[o]d create-3sgm pst-G[o]d create-def.f-heaven
‘God created him.’ ‘God created heaven.’

Similarly, all combinations of incorporation and case marking are possible for
lexical noun phrase core arguments.

Table 5: Co-occurrence of case-marking and incorporation for lexical NP arguments

Case-marked P Incorporated P
Case-marked A Ex. 11 a-u-el-ôni nče-ni-ouidai
pst-3pl-throw-stone [Link]-Jew
Incorporated A Ex. 5 Ex. 8

All in all, Coptic can be described as having both Differential Subject Marking and
Differential Object Marking, which is in itself rare. Coptic Differential Argument
Marking is also of a relatively rare type, in which overt case marking alternates
with argument incorporation (Grossman & Iemmolo 2014+). Differential Subject
Marking in Coptic is triggered by information structure: in the Bohairic dialect,
preverbal subjects tend to be contrastive or shifted topics; incorporated subjects
tend to be relatively “inert” in terms of topicality, with little backward or forward
topicality in texts, or otherwise globally accessible referents (God, evening,
morning); postverbal subjects tend to be highly accessible (Ariel 1990), already
active in the discourse.10 This last observation is discussed in the next section.

10  For details and argumentation, see Zakrzewska (2006); Shisha-Halevy (1986, 2007); and
Grossman & Iemmolo (2014+). Loprieno (2000) proposes that postverbal subjects are rhematic,
and Reintges (2004) considers that they are focal. In fact, these views are not necessarily con-
tradictory, since Reintges (2004) deals with Sahidic, in which postverbal subjects are more fre-
quently new referents. This, however, is very rare in Bohairic.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
212   Eitan Grossman

2.6 Linear order

Coptic is usually characterized as having basic SVO order, although this descrip-
tion is somewhat controversial. While a description of word order in Coptic is well
beyond the scope of this paper, some relevant facts should be mentioned.
1. The order of indexes on the verb is distinctive: an A index always precedes a
P index. The order is always A-Vlex-P.11
2. The order of incorporated core arguments is also distinctive; it is always
A-Vlex-P.
3. Lexical A, no matter its position, almost always precedes lexical P. One type
of exception is when the lexical object is preverbal and the lexical subject is
postverbal, e.g.

(17) ⲛⲁⲓ ⲇⲉ ⲁϥϫⲟⲧⲟⲩ ⲛϫⲉⲓⲏⲥ

nai=de a-f-čot-ou nče-iê[sou]s
[Link]=conn pst-3sgm-say-3pl nom-Jesus
‘As for these things, Jesus said them…’ (Early Bohairic, John 12: 36).

This construction type, however, is very rare in discourse.

There are also rare examples in which postverbal lexical P precedes lexical A:

(18) ⲁϥϫⲓ ⲇⲉ ⲛⲛⲓⲱⲓⲕ ⲛϫⲉⲓⲏⲥ

a-f-či=de n-ni-ôik nče-iê[sou]s
pst-3sgm-take=ptcl [Link]-bread nom-Je[su]s
‘Then Jesus took the loaves of bread’ (Early Bohairic, John 6: 11).

However, all in all, lexical A overwhelmingly tends to precede lexical P.

Linear order alternations in Coptic, as in many languages, are motivated by
information-structural factors, and as such, linear order is a “coding means” in
the sense of Frajzyngier & Shay (2003) and Frajzyngier (2011). As noted above,
preverbal arguments are typically contrastive or shifted topics, although, they
can occasionally be focal. Postverbal A/S arguments in Coptic, at least in some
dialects, e.g., Bohairic, are generally identifiable/accessible (Zakrzewska 2006;
Grossman & Iemmolo 2014+).
For example, in (18), ‘Cain’ occurs in the first clause; in the second clause,
‘Cain’ is already accessible, and occurs postverbally with nom-marking.

11  In terms of linear order, this construction has nominative-accusative alignment, even if the
person indexes themselves are neutrally aligned.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   213

(19) ⲁⲡϭⲥ ⲫϯ ⲭⲁⲟⲩⲙⲏⲓⲛⲓ ⲛⲕⲁⲓⲛ

a-pc[oi]s p [nou]t k a-ou-mêini
h i h
n-kain
[Link] G[o]d [Link]-sign acc-Cain
‘The Lord G[o]d marked Cain.’

ⲁϥⲓ ⲇⲉ ⲛϫⲉⲕⲁⲓⲛ ⲉⲃⲟⲗ ϩⲁⲡϩⲟ ⲙⲫϯ

a-f-i=de nče-kain ebol ha-p-ho m-ph[nou]ti
pst-3sgm-come=conn nom-Cain out from-def.m-face of-G[o]d
‘(and) Cain came away from the Lord’s presence’ (Bohairic, Genesis 4:
15–16).

In the Biblical story of Jonah, the ship bearing the runaway is rocked by a storm,
which frightens the sailors. Jonah admits that the storm is his fault.

(20) [ⲉ]ⲑⲃⲏⲧ ϥϣⲟⲡ ⲛϫⲉⲡⲁⲓϩⲱⲓⲙⲓ

[e]thbêt f-šop nče-pai-hôimi
[b]ecause.1sg 3sgm-exist nom-dem.m-wave
‘Because of me this wave exists’ (Early Bohairic, Jonah 1:12).

The nom-marked subject is the current discourse topic, and as such, is identifi-
able.
In the following example, Jesus and an official are discussing the latter’s ill
son, and Jesus says:

(21) ϥⲟⲛϧ ⲛϫⲉⲡⲉⲕϣⲉⲣⲓ

f-onx nče-pek-šêri
3sgm-alive nom-your-son
‘Your son is alive’ (Early Bohairic, John 4:50).

Highly identifiable/accessible postverbal arguments are sometimes called anti-

topics (Chafe 1976). Antitopics, according to Lambrecht (1981), tend to have the
following cluster of properties: (1) postverbal position; (2) indexed via pronomi-
nal affixes/clitics; (3) case-marked and integrated into clausal syntax, i.e., are not
clause-external; (4) identifiable (textually, frame-evoked, inferable, following
Prince 1981), i.e., they do not introduce new topics. Postverbal subjects in Coptic,
at least in the dialect discussed here, have all of these properties.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
214   Eitan Grossman

3 Frequency
To date, there are no extensive studies of the relative frequency of the various
constructions in discourse, other than Zakrzewska (2006), which provides raw
frequency data for around 700 narrative clauses in a single corpus of later (ca. 9th
century) Bohairic.

Table 6: Subject encoding in later Bohairic

A/S-incorporation Preverbal A/S Postverbal A/S (nom)

31.5% 25.7% 42.8%

Note that nearly 70% of clauses have non-incorporated subjects, and nearly half
of the tokens have postverbal nom-marked subjects.
Two small samples of Early Bohairic were examined. In the first, taken from
Genesis (Papyrus Bodmer III), all clauses were analyzed, with the following
results:

Table 7: Subject encoding in Genesis 1:1–4:3 in all clauses

Incorporation Preverbal Postverbal Index only Other

Number 68 12 29 79 80
% 25.373 4.478 10.821 29.477 29.851

For clauses with lexical subjects, one sees that incorporated subjects amount to
nearly 70% of all tokens, with preverbal and postverbal subjects far less frequent.

Table 8: Subject encoding in Genesis 1:1–4:3 in clauses with lexical S/A

Incorporation Preverbal Postverbal Total

Number 68 12 29 109
% 62.385 11.009 26.606 100

However, this has to do with the nature of the text examined: in the first 200
clauses or so of Genesis, God is the only actor, so there is little need for formal
devices that mark shifted or contrastive topics (e.g., left-dislocation), or that keep
track of alternating subject referents (e.g., postverbal nom-marked antitopics).
A second sample, from the Gospel of John, was looked at, first counting all
clauses, then only clauses with lexical subjects.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   215

Table 9: Subject encoding in the Gospel of John 18:1–19:16 in all clauses

Incorporation Preverbal Postverbal Index only Other

Number 10 20 36 71 99
% 4.237 8.474 15.255 30.085 41.949

Table 10: Subject encoding in the Gospel of John 18:1–19:16 in clauses with lexical S/A

Incorporation Preverbal Postverbal Total

Number 10 20 36 66
% 15.152 30.303 54.545 100

In John (18:1–19:16), the frequencies resemble those found by Zakrzewska (2006)

for later Bohairic, with postverbal subjects even more frequent (more than half of
tokens of clauses with lexical subjects). This is probably due to the high density of
actors (e.g., Jesus, Pilate, the Jews, various disciples, the priests), who have to be
kept track of in narrative by formal devices. Importantly, all of the actors are iden-
tifiable/accessible in this chunk of text, which accounts for the high frequency of
postverbal nom-marked subjects as opposed to preverbal subjects.

4 Does Coptic have a “marked-nominative”

system?
According to König (2008; 2009), the vast majority of African languages with
“no case before the verb” have either ergative or “marked nominative” coding.
Marked nominative constructions are those in which S and A are encoded by the
same means, differing from that of P, but unlike other nominative-accusative con-
structions, the accusative is the morphologically and functionally “unmarked”
form. In this context, “unmarked” means that accusatives are less morphosyn-
tactically complex than nominatives (“morphologically unmarked”), on the one
hand, and that accusatives are used in more contexts (“functionally unmarked”).
König attributes special importance to the citation form, which in marked nomi-
native systems is the accusative.
Coptic is formally similar to marked nominative systems, in that the nomi-
native is the most functionally restricted case, both syntactically and pragmati-
cally, and tends to occur postverbally in many marked nominative systems. On
the other hand, it differs from most marked nominative systems in that both the

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
216   Eitan Grossman

nom and the acc are restricted, in terms of function, relative to the zero-marked
(unmarked or “bare”) form. Moreover, neither the nom nor the acc is the citation
form (“absolute” or “designative” case, Creissels 2009), like Japanese, Korean,
and Kanuri. Of course, another major difference is the tripartite distinction
between preverbal, postverbal, and incorporated subjects, which is atypical of
marked nominative systems. The following Table gives an idea of the functions
of the three forms:

Table 11: Functions of case markers in Coptic

Case Function
nom (nče-) (a) post-verbal A/S
acc (n-) (b) post-verbal P, some adverbials, secondary predicates
unmarked (c) citation form
(d) vocative
(e) following prepositions
(f) incorporated S/A/P
(g) preverbal S/A/P
(h) nominal predications
(i) following derivational affixes

As such, Coptic cannot be described as having a marked nominative system, but

rather as having an “uncommon pattern of accusative core marking”, namely
“marked S/A vs. marked P” (Creissels 2009: 453). On the other hand, Coptic does
share with many languages with marked nominative systems the property of a
basically pragmatic or information-structural basis for the occurrence of case
markers. However, this is not exclusive to marked nominative systems: in some
languages, postverbal subjects tend to refer to identifiable or accessible entities
in discourse,12 while preverbal subjects have a marked information-structural
status, e.g., focus, on the one hand, or contrastive or shifted topic, on the other.
What is particular about the Coptic system, and appears to be a cross-linguistic
rarum is the fact that postverbal subjects have a strong association with antitopic
status and are obligatorily coded by a highly grammaticalized case marker.

12  Such languages include Syriac Aramaic (Goldenberg 1983), Biblical Hebrew (Givón 1977),
Ojibwa (Tomlin & Rhodes 1992), Mohawk (Mithun 1996), French (Lambrecht 1981), Chamorro
(Cooreman 1992), and Sandawe (Eaton 2010). However, to the best of my knowledge there are no
cross-linguistic studies of antitopics.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   217

It is likely that an explanation for this association between nominative

marking and highly accessible lexical noun phrase referents is another instan-
tiation of a general trend: overt case marking, at least at early stages of gram-
maticalization, marks unexpected associations between grammatical role and
information-structural properties of referents. For example, Iemmolo (2010)
argues that differential object marking tends to arise in the context of topical
P, since direct objects tend to be new (or “focal”). Similarly, numerous studies
of “optional” ergative marking point out that “optional” ergative markers are
strongly associated with focal A (e.g., Hyslop 2010; McGregor 2006; Verstraete
2010). This makes sense, since A referents overwhelmingly tend to be highly
accessible (Du Bois 1987). In the present context, lexical noun phrases tend to
refer to low-accessibility referents, so the occurrence of highly accessible refer-
ents as lexical noun phrases would be an unexpected association, and therefore
would be prone to being marked overtly. While a number of languages exploit
word order to mark this unexpected association (e.g., French and Ojibwa), Coptic
seems to be unusual in using case marking.

5 A brief word on the diachrony of Differential

Subject Marking in Coptic13
Of the three constructions discussed here – subject incorporation, preverbal sub-
jects, and postverbal nom-marked subjects – the subject incorporation construc-
tion represents the most frequent construction of Earlier Egyptian, a largely VS
language. In Earlier Egyptian, person markers and lexical noun phrases were in
complementary distribution within the same clause, attaching directly to verb
stems. As such, person markers are “pronominal” in the narrow sense, or “pro-
indexes” in the sense of Haspelmath (2013), i.e., “person indexes” that cannot
co-occur with conominals in the same clause.
In Later Egyptian, there began to emerge numerous periphrastic and aux-
iliary constructions, to which subject expressions attached, followed by lexical

13  Differential Object Marking in Egyptian-Coptic, which began to emerge in Late Egyptian,
developed from the convergence of a number of constructions, including an antiapplicative
construction and a partitive-object construction. The accusative marker itself developed from
the highly polysemous preposition m, which codes instrument, source, location, and a range of
other functions. See Winand (2014, in this volume) for further details of the diachrony of the ac-
cusative marker from Late Egyptian to Coptic.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
218   Eitan Grossman

verbs. Coptic TAM auxiliaries represent, for the most part, highly grammatical-
ized descendants of these auxiliaries.

Demotic ( Johnson 1976: 97)

(22) [Link]-msḥ ṯɜy rmt n-dmy
[Link]-crocodile catch man mod-town
‘A crocodile does not/cannot catch a local man.’

(23) [Link]-f-ḫpr
[Link]-3sgm-exist
‘It does not/cannot happen.’
Coptic (Shisha-Halevy 1988: 97)
(24) ⲙⲉⲣⲉⲡⲉϥϩⲏⲧ ϩⲓⲛⲏⲃ ⲛⲟⲩⲙⲉ
mere-pef-hêt hinêb noume
[Link]-his-heart sleep really
‘His heart cannot really sleep.’

(25) ⲙⲉⲕϩⲱⲡ ⲉⲣⲟⲟⲩ

me-k-hôp ero-ou
[Link]-2sgm-hide all-3pl
‘You cannot hide from them.’

As a result of these grammaticalization processes, lexical noun phrase subjects

and person indexes became “trapped” within the verbal syntagm, rather than at
its rightmost edge.
Since Earlier Egyptian, preverbal lexical subjects could be right-dislocated or
topicalized, as an alternative order. However, there is no comparative frequency
data for preverbal subjects in Earlier and Later Egyptian.
In contrast to the previous two constructions, the postverbal nominative
marker is largely an innovation of Coptic, with vanishingly few examples in the
previous stage of the language, Demotic. Another marker, written variously in, n,
or m, was relatively infrequent (Mattha 1947). The frequency of postverbal sub-
jects is vastly greater in Coptic than in any previous stage of the language.
The implications of these processes for understanding the history of the
encoding of grammatical relations in Egyptian are significant. In a recent decon-
struction of the concept of agreement, Haspelmath (2013) argues that the term
“index” is more appropriate for describing bound person markers. The proposed
typology of bound person markers distinguishes between three main types:

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   219

1. Pro-indexes: indexes that cannot co-occur with conominals

2. Cross-indexes: indexes with optional conominals
3. Gramm-indexes: indexes with obligatory conominals

In the latest stages of Egyptian-Coptic, one witnesses a sharp increase in the co-
occurrence of person indexes and conominals, or in other words, a shift from
pro-indexes to cross-indexes, due to a rise in the frequency of preverbal and post-
verbal lexical arguments. However, since indexes can occur without conominals
in all stages of Coptic (ca. 30% in the corpora checked here), there is no stage of
the Egyptian-Coptic language in which one can speak of gramm-indexes.
Nonetheless, one observes a diachronic tendency for index + conominal as a
discourse preference or “soft constraint”, realized by the high text frequency of
such constructions vis-à-vis subject incorporation.

6 No case before the verb in northeast Africa

Despite Africa’s reputation as a caseless continent, König (2008; 2009) has
argued that many African languages should in fact be described as case lan-
guages, whether case is realized by bound forms, free forms, or tone. This charac-
terization assumes a broad definition of the comparative concept “case marker”,
similar to that adopted here
Northeast Africa, with Berber as a possible geographical outlier, is character-
ized by a robust generalization: preverbal core arguments show no case distinc-
tions; the form that occurs preverbally is “always in one case form only, namely,
the morphologically most unmarked one” (König 2008: 240). This generalization
holds regardless of alignment type, since it holds for both “marked nominative”
(e.g., Turkana, Nilo-Saharan) and ergative systems (e.g., Shilluk, Nilo-Saharan).
In Päri (West Nilotic, Nilo-Saharan), which has a split ergative/ marked-nom-
inative case system, pre-verbal A and P are in the absolute form, which does not
have an overt case marker. On the other hand, post-verbal A and P show overt
case markers.

(27) Päri (West Nilotic, Nilo-Saharan; Andersen 1988)

(a) rìŋó ŋôl ùbúrr-ì ŋɔ́l-ɔ̀
meat cut Ubur-erg cut-suf
‘Ubur will cut the meat.’

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
220   Eitan Grossman

(b) ùbúr ŋùt-ò kí rìŋó

Ubur [Link]-suf obl meat
‘Ubur will cut the meat.’

König’s generalization trivially does not apply to the majority of accusative lan-
guages of northeastern African languages in her sample, since all but one of her
accusative languages are verb-final; this fact renders the generalization irrele-
vant, since if the “no case before the verb” generalization were to hold for a verb-
final language, it would not be considered a case language.
In many of the languages in König’s sample, linear order is motivated by the
information structural properties of referents in discourse. In Dinka (West Nilotic,
Nilo-Saharan, Andersen 1991; 2002), Nandi (Kalenjin, South Nilotic, Nilo-Saha-
ran, König 2008: 260), Chai (Southeast Surmic, Nilo-Saharan, König 2008: 247),
and Datooga (South Nilotic, Nilo-Saharan, König 2008: 261), preverbal arguments
are topical. On the other hand, in Toposa (East Nilotic, Nilo-Saharan), Maa (East
Nilotic, Nilo-Saharan, König 2008: 262; Payne 2013), and Tennet (König 2008:
262), preverbal arguments are generally focal. In Turkana (Nilo-Saharan), prever-
bal position can accommodate both topical and focal arguments (Dimmendaal
1983: 408; König 2008: 259). In Shilluk (Nilo-Saharan), it is postverbal arguments
that are focal (König 2008: 243).
Kanuri (Saharan, Bondarev et al. 2011) Differential Subject Marking is moti-
vated by the interaction of semantic and pragmatic features. The “nominative”
marker –ye ”is admissible providing they represent addressee-old information
and occur with either transitive verbs/clauses or (if human) intransitive active
verbs/clauses” (Bondarev et al. 2011).14 This description is similar to that pro-
posed for Coptic in this article.

7 Concluding remarks
As mentioned in the introduction, the goal of this article is a modest one: to draw
attention to a grammatical feature that has hitherto gone unmentioned in most
descriptions of Coptic: no case before the verb, obligatory case after the verb,
although the Coptic facts would better be described as 'obligatory case after the
verb, no case anywhere else'. From an areal point of view, this is not unexpected:
Coptic closes the geographical gap, so to speak, between the main cluster of

14  I would like to thank Denis Creissels for pointing out to me the relevance of Kanuri, and
Dmitry Bondarev for sharing his work on the language with me.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   221

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
222   Eitan Grossman

languages in which preverbal core arguments are not overtly flagged, in north-
eastern Africa, and North African Berber languages, which have until now been
considered a geographical outlier. This can be seen from König’s map of “no case
before the verb” languages in Africa (2008: 274).
Since this feature is so striking, it is interesting to consider why it has not
been explicitly mentioned in descriptions. It may be due, at least in part, to the
fact that many descriptions of Coptic do not gloss examples, so there is little need
to decide on labels for descriptive categories. It probably also has something to do
with the fact that Coptic is not typically described as a case language.
What is interesting about the Coptic case system is not the association of
postverbal subjects with antitopic status, since this seems to be cross-linguisti-
cally well attested. Nor is it its three-way case system with overt nominative and
accusative, neither of which is the citation form; similar systems are attested,
albeit rarely, in, e.g., Korean, Japanese, and Kanuri. Nor is it the “no case before
the verb, obligatory case after the verb” feature, which is typical of the broad area
in which Coptic was spoken. Rather, it is the particular constellation of these fea-
tures: an overt nominative marker which differs from both the accusative and the
citation form, and which obligatorily marks highly accessible postverbal subjects.

Abbreviations
The glossing conventions here are in accordance with the Leipzig Glossing Rules,
and the transliteration follows the Leipzig-Jerusalem transliteration system
(Grossman & Haspelmath 2014, in this volume). The following are language-spe-
cific glosses.

aor aorist verb form

conn connecting particle
mod modifier marker

8 References
Aikhenvald, Alexandra. 2007. Typological distinctions in word-formation. In: Shopen, Timothy
(ed.), Language Typology and Syntactic Description. Vol. 3, Grammatical Categories and
the Lexicon. 2nd ed. Cambridge: Cambridge University Press.
Andersen, Torben. 1988. Ergativity in Pari, a Nilotic OVS language. Lingua 75: 289–324.
Andersen, Torben. 1991. Subject and topic in Dinka. Studies in Language 15: 265–94.
Andersen, Torben. 2002. Case inflection and nominal head marking in Dinka. Journal of African
Languages and Linguistics 23: 1–30.
Ariel, Mira. 1990. Accessing Noun-Phrase Antecedents. London/New York: Routledge.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   223

Arkadiev, Peter. fc. The Kabyle “state” distinction: dependent-marking after all? A commentary
on Mettouchi & Frajzyngier 2013. Ms. available [Link]
THE_BERBER_STATE_DISTINCTION_DEPENDENT_MARKING_AFTER_ALL_A_COMMENTARY_
ON_METTOUCHI_and_FRAJZYNGIER_2013
Bondarev, Dmitry, Jaggar, Philip J., Löhr, Doris & Tijani, Abba I. 2011. Differential subject
marking in Kanuri: agentivity, pragmatics, and split-intransitive. In: Löhr, Doris, Rothmaler,
Eva & Ziegelmeyer, Georg (eds.), Kanuri, Borno and Beyond: Current Studies on the Lake
Chad Region. (Topics in Interdisciplinary African Studies 22). Köln: Rüdiger Köppe, 27–57.
Chafe, Wallace. 1976. Givenness, contrastiveness, definiteness, subjects, topics and point of
view. In: Li, Charles N. (ed.), Subject and topic. New York: Academic Press, 25–55.
Cooreman, Ann. 1992. The pragmatics of word order variation in Chamorro narrative text. In:
Payne, Doris L. (ed.), Pragmatics of Word Order Flexibility. Amsterdam: John Benjamins,
243–263.
Creissels, Denis. 2009. Uncommon patterns of core term marking and case terminology. Lingua
119: 445–459.
Dimmendaal, Gerrit. 1983. Turkana as a verb-initial language. Journal of African Languages and
Linguistics 5: 17–44.
Dixon, Robert M.W. 1988. A Grammar of Boumaa Fijian. Chicago/London: University of Chicago
Press.
Du Bois, John. 1987. The Discourse Basis of Ergativity. Language 63: 805–855.
Eaton, Helen. 2010. Information structure marking in Sandawe texts. In: Fiedler, Ines &
Schwarz, Anne (eds.), The Expression of Information Structure: A documentation of its
diversity across Africa. Amsterdam/Philadelphia: John Benjamins, 1–34.
Engsheden, Åke. 2008. Differential Object Marking in Coptic. In: Josephson, F. & Söhrman, I.
(eds.), Forms and Functions: Aspect, Tense, Mood, Diathesis and Valency. Amsterdam:
John Benjamins, 323–344.
Frajzyngier, Zygmunt. 2011. Les functions de l'ordre lineaire. Bulletin de la Société Linguistique
de Paris CV.1: 7–37.
Frajzyngier, Zygmunt & Shay, Erin. 2003. Explaining language structure through systems
interaction. Amsterdam/Philadelphia: John Benjamins.
Funk, Wolf-Peter. 1988. Dialects wanting homes: a numerical approach to the early varieties
of Coptic. In: Fisiak, Jacek (ed.), Historical dialectology: Regional and Social. Berlin/New
York/Amsterdam: Mouton de Gruyter, 149–192.
Givón, Talmy. 1977. The Drift from VSO to SVO in Biblical Hebrew: the Pragmatics of
Tense-Aspect. In: Li, Charles N. (ed.), Mechanisms of Syntactic Change. Austin: University
of Texas Press, 184–254.
Good, Jeff. 2010. Topic and focus fields in Naki. In: Fiedler, Ines & Schwarz, Anne (eds.),
The Expression of Information Structure: A documentation of its diversity across Africa.
Amsterdam/Philadelphia: John Benjamins, 35–67.
Goldenberg, Gideon. 1983. On Syriac Sentence Structure. In: Sokoloff, Michael (ed.), Arameans,
Aramaic, and the Aramaic Literary Tradition. Ramat Gan: Bar-Ilan University Press, 97–140.
Grossman, Eitan. 2009a. Studies in Early Bohairic Syntax. PhD thesis. Hebrew University of
Jerusalem.
Grossman, Eitan. 2009b. Periphrastic Perfects in the Coptic Dialects: A Case Study in Grammati-
calization. Lingua Aegyptia 17: 81–118.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
224   Eitan Grossman

Grossman, Eitan & Iemmolo, Giorgio. 2014+. A rare case of differential marking on A/S: the
case of Coptic. Paper presented at the Association for Linguistic Typology 10th Biennial
Conference, Leipzig 2013.
Haspelmath, Martin. 2008. Terminology of Case. In: Malchukov, Andrej & Spencer, Andrew
(eds.), The Oxford Handbook of Case. Oxford: Oxford University Press, 505–517.
Haspelmath, Martin. 2011a. The indeterminacy of word segmentation and the nature of
morphology and syntax. Folia Linguistica 45.2: 31–80.
Haspelmath, Martin. 2011b. On S, A, P, T, and R as comparative concepts for alignment typology.
Linguistic Typology 15: 535–567.
Haspelmath, Martin. 2013. Argument indexing: a conceptual framework for the syntactic status
of bound person forms. In: Bakker, Dik & Haspelmath, Martin (eds.), Languages across
boundaries: Studies in memory of Anna Siewierska. Berlin: Mouton de Gruyter.
Haspelmath, Martin. 2014. A grammatical overview of Egyptian and Coptic. This volume.
Hyslop, Gwendolyn. 2010. Kurtop case: the pragmatic ergative and beyond. Linguistics of the
Tibeto-Burman area 33.1: 1–40.
Iemmolo, Giorgio. 2010. Topicality and differential object marking: evidence from Romance and
beyond. Studies in Language 34.2: 239–272.
Johnson, Janet H. 1976. The Demotic Verbal System. (Studies in Ancient Oriental Civilization 38).
Chicago: Oriental Institute.
König, Christa. 2008. Case in Africa. Oxford: Oxford University Press.
König, Christa. 2009. Marked Nominatives. In: Malchukov, Andrej & Spencer, Andrew (eds.),
The Oxford Handbook of Case. Oxford: Oxford University Press, 534–548.
Lambrecht, Knud. 1981. Topic, Antitopic, and Verb Agreement in Non-Standard French.
Amsterdam: John Benjamins.
Layton, Bentley. 2004. A Coptic Grammar. Sahidic Dialect. Wiesbaden: Harrassowitz.
Lazard, Gilbert. 2002. Transitivity revisited as an example of a more strict approach in
typological research. Folia Linguistica 36.3–4: 141–190.
Loprieno, Antonio. 2000. From VSO to SVO? Word Order and Rear Extraposition in Coptic. In:
Sornicola, Rosanna, Poppe, Erich & Shisha-Halevy, Ariel (eds.), Stability, Variation and
Change of Word Order Patterns over Time. Amsterdam/Philadelphia: John Benjamins,
23–39.
Mattha, Girgis. 1947. The Origin of the Explanatory Particles n and nci/nče. Bulletin de l’Institut
Français d’Archaéologie Orientale 45: 61–64.
McGregor, William. 2006. Focal and optional ergative marking in Warrwa (Kimberley, Western
Australia). Lingua 116: 393–423.
Mettouchi, Amina & Frajzyngier, Zygmunt. 2013. A previously unrecognized typological
category: The state distinction in Kabyle (Berber). Linguistic Typology 17: 1–30.
Mithun, Marianne. 1996. The Mohawk language. In: Maurais, Jacques (ed.), Quebec's
Aboriginal Languages: History, Planning and Development. Clevedon/Avon, England:
Multilingual Matters, 159–173.
Payne, Doris L. 1992. Nonidentifiable information and pragmatic order rules in ‘O’odham. In:
Payne, Doris L. (ed.), Pragmatics of Word Order Flexibility. Amsterdam: John Benjamins,
137–167.
Payne, Doris L. 2013. Grammatical Relations and Alignment. In: Luraghi, Silvia & Parodi,
Claudia (eds.), Bloomsbury Companion to Syntax. London/New Delhi/New York/Sydney:
Bloomsbury, 218–239.
Polotsky, Hans Jakob. 1960. The Coptic Conjugation System. Orientalia 29: 392–422.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
 No case before the verb   225

Prince, Ellen. 1981. Toward a Taxonomy of Given-New Information. In: Cole, Peter (ed.), Radical
Pragmatics. New York: Academic Press, 223–254.
Reintges, Chris. 2004. Coptic Egyptian (Sahidic Dialect). Köln: Rüdiger Köppe.
Shisha-Halevy, Ariel. 1986. Coptic Grammatcial Categories. Structural Studies in the Syntax of
Shenoutean Sahidic. Roma: Pontificium Institutum Biblicum.
Shisha-Halevy, Ariel. 1988. Coptic Grammatical Chrestomathy. Leuven: Peeters.
Shisha-Halevy, Ariel. 2007. Topics in Coptic Syntax: Structural Studies in the Bohairic Dialect.
Leuven/Paris/Dudley, MA: Peeters.
Siewierska, Anna & Bakker, Dik. 2008. Case and alternative strategies. In: Malchukov, Andrej &
Spencer, Andrew (eds.), Handbook of Case. Oxford: Oxford University Press, 290–303.
Tomlin, Russell S. & Rhodes, Richard A. 1992. Information distribution in Ojibwa. In: Payne,
Doris L. (ed.), Pragmatics of Word Order Flexibility. Amsterdam: John Benjamins, 117–135.
Verstraete, Jean-Christophe. 2010. Animacy and information structure in the system of ergative
marking in Umpithamu. Lingua 120: 1637–1651.
Winand, Jean. 2014. The Oblique Expression of the Object in Ancient Egyptian. This volume.
Zakrewska, Ewa 2006. The Hero, the Villain and the Mob: Topicality and Focality in Bohairic
Narrative Discourse. Lingua Aegyptia 14: 325–346.

Brought to you by | Cambridge University Library

Authenticated
Download Date | 11/24/15 7:55 AM
Brought to you by | Cambridge University Library
Authenticated
Download Date | 11/24/15 7:55 AM