Pro-drop in Hebrew: a summary

— this is a draft encyclopedia entry —

Pro-drop is an abbreviation of “pronoun dropping.” It describes a feature of some languages that do not require an overt argument, especially a subject, to be present in a clause. That is, whereas English is not a pro-drop language and thus requires a subject noun or pronoun in a finite verbal clause like He has spoken, in Italian the overt subject may be “dropped,” Ha parlato ‘(He) has spoken’.

Languages that allow pro-drop fall into three general categories (see Huang 1984; also Dryer 2008): those that allow pro-drop only in restricted environments (e.g., English, where a subject can be dropped only in non-tensed clauses, such John preferred __ seeing Mary vs. John preferred that he see Mary); those that allow pro-drop in most subject positions, but not the object position (e.g., Italian, Spanish); and languages that allow both subject and object pro-drop (e.g., Chinese, Japanese). Biblical Hebrew and Modern Hebrew fall into the third category: they allow pro-drop in both subject and object positions. The biblical example in (1) and modern example in (2) illustrate null subjects in Hebrew.

(1) וַתִּקַּ֥ח מִפִּרְי֖וֹ וַתֹּאכַ֑ל וַתִּתֵּ֧ן גַּם־לְאִישָׁ֛הּ עִמָּ֖הּ וַיֹּאכַֽל

wat-tiqqaḥ mip-piryō wat-tōʾḵal wat-tittēn gam lǝʾīšå̄h ʿimmå̄h way-yōʾḵal

‘and pro (Eve) took from its fruit and pro (she) ate (it) and pro (she) gave (it) to her husband who was with her and pro (he) ate (it)’ (Gen 3.6)


(2) ‏(אני) אכלתי את התפוח

(ʾani) ʾaxalti ʾet ha-tapuax̱

‘(I) pro ate the apple’ (Borer 1986:376).

As of yet, there has been no focused study on subject pro-drop in most stages of pre-modern Hebrew. The lone exception is Naudé’s work on null subjects in Qumran Hebrew (Naudé 1991, 1996, 2001). Naudé also compares the syntactic distribution of null subjects and overt subjects in Qumran Hebrew, Biblical Aramaic, and Mishnaic Hebrew, with a final discussion of some differences in Biblical Hebrew. Pro-drop in Modern Hebrew, in contrast, has been a focus of numerous studies (see, among others, Borer 1986, Shlonsky 1987, 2009, Doron 1988, Ritter 1995, Vainikka and Levy 1999, Gutman 1999, 2004).

Though a full scale study of pro-drop in any stage of pre-modern is lacking, it has long been observed that the subject pronoun is not obligatory and also that an overt pronoun signals ’emphasis’ (i.e., Topic or Focus); see GKC:§135, Joüon and Muraoka 2006:§146; Heimerdinger 1999, Shimasaki 2002.

One feature that distinguishes Modern Hebrew is that the null subject may be used only with first and second-person past and future verbs; with third-person past and future verbs in main clauses and all present tense verbs the subject pronoun is obligatory (3).

(3) ‏הוא אכל את התפוח vs. אכל את התפוח*

huʾ ʾaxal ʾet ha-tapuax̱ *ʾaxal ʾet ha-tapuax̱

‘he ate the apple’

For this reason, Modern Hebrew is sometimes called partial pro-drop (see, for example, Shlonsky 2009). Although pro-drop is most often used to refer to the phenomenon of a null subject, languages falling into the third category of pro-drop demonstrate that the null pronoun can also appear in the object position (Rizzi 1986). This is certainly so for Biblical (4) and Modern (5) Hebrew.

(4) ‏ וַתִּקַּ֥ח מִפִּרְי֖וֹ וַתֹּאכַ֑ל וַתִּתֵּ֧ן גַּם־לְאִישָׁ֛הּ עִמָּ֖הּ וַיֹּאכַֽל

wat-tiqqaḥ mip-piryō wat-tōʾḵal wat-tittēn gam lǝʾīšå̄h ʿimmå̄h way-yōʾḵal

‘and pro (Eve) took from its fruit and pro (she) ate pro (it) and pro (she) gave pro (it) to her husband who was with her and pro (he) ate pro (it)’ (Gen 3.6)


(5) ‏תקנה את התפוח ותתן לי

tiqne ʾet ha-tapuax̱ vetiten li

‘buy the apple and give pro (it) to me’

Both examples in (4) and (5) illustrate that the null object is dropped only when the antecedent is accessible within the discourse (see Gutman 1999:126-254 (chp. 3)). In both (4) and (5) the antecedents of the null object pronouns (fruit and apple, respectively) are explicitly identified in the preceding clauses. (See Creason 1991 for the only study of null object anaphora in Biblical Hebrew.)

Within generative theory, the “dropped” pronoun has been identified as a type of empty or null category — a phonologically empty but syntactically real pronoun — and is referred to as pro (read as “little pro”). Early on in the generative approach to pro-drop, the phenomenon was associated with the nature of verbal inflection, since in many pro-drop languages the finite verb is inflected with morphologically rich affixes (i.e., the verbal affixes are portmanteau morphs, carrying a bundle of person, number, and gender features). While the agreement affixes in these languages (such as Italian, Spanish, Arabic, Hebrew) may aid in the identification of the null subject pro, morphological agreement between the verb and subject is clearly not a prerequisite for pro-drop, since Chinese and Japanese, for example, are pro-drop languages without verbal agreement features.


Borer, Hagit. 1986. “I-Subjects.” Linguistic Inquiry 17 (3):375-416.

Creason, Stuart. 1991. “Discourse Constraints on Null Complements in Biblical Hebrew.” University of Chicago Working Papers in Linguistics 7:18-47.

Doron, Edit. 1988. “On the Complementarity of Subject and Subject-verb Agreement.” Agreement in Natural Language: Approaches, Theories, Descriptions, ed. M. Marlow and C. Ferguson, 202-18. Stanford: CSLT.

Dryer, Matthew S. 2008. “Expression of Pronominal Subjects.” The World Atlas of Language Structures Online, ed. Haspelmath, Martin, Matthew S. Dryer, David Gil, and Bernard Comrie, chp. 101. Munich: Max Planck Digital Library. (Accessed on 2010-05-28.)

Gutman, Eynat. 1999. Null Subjects: A Theory of Syntactic and Discourse-Identification, Linguistics, University of Delaware.

———. 2004. “Third Person Null Subjects in Hebrew, Finnish and Rumanian: An Accessibility-Theoretic Account.” Journal of Linguistics 40:463-90.

Huang, James C.-T. 1984. “On the Distribution and Reference of Empty Pronouns.” Linguistic Inquiry 15 (4):531-74.

Heimerdinger, Jean-Marc. 1999. Topic, Focus and Foreground in Ancient Hebrew Narratives. JSOTSupp 295. Sheffield: Sheffield Academic Press.

Joüon, Paul, and Takamitsu Muraoka. 2006. A Grammar of Biblical Hebrew. Rev. ed. Subsidia Biblica 27. Rome: Pontifical Biblical Institute.

[GKC] Kautzsch, Emil. 1910. Gesenius’ Hebrew Grammar. Trans. A. E. Cowley. Second English ed. Oxford: Clarendon Press.

Naudé, Jacobus A. 1991. “Qumran Hebrew as a Null Subject Language.” South African Journal of Linguistics 9 (4):119-25.

———. 1993. “On Subject Pronoun and Subject Noun Asymmetry: A Preliminary Survey of Northwest Semitic.” South African Journal of Linguistics 11 (1):17-28.

———. 1996. Independent Personal Pronouns in Qumran Hebrew Syntax. Ph.D. thesis, University of the Free State, South Africa.

———. 2001. “The Distribution of Independent Personal Pronouns in Qumran Hebrew.” Journal of Northwest Semitic Languages 27 (2):91-112.

Ritter, Elizabeth. 1995. “On the Syntactic Category of Pronouns and Agreement.” Natural Language and Linguistic Theory 13 (3):405-43.

Rizzi, Luigi. 1986. “Null Objects in Italian and the Theory of pro.” Linguistic Inquiry 17:501-57.

Shimasaki, Katsuomi. 2002. Focus Structure in Biblical Hebrew: A Study of Word Order and Information Structure. Bethesda, MD: CDL Press.

Shlonsky, Ur. 1987. Null and Displaced Subjects. Ph.D. thesis, Massachusetts Institute of Technology, Cambridge, MA.

———. 2009. “Hebrew as a Partial Null-Subject Language.” Studia Linguistica 63 (1):133-57.

Vainikka, Anne, and Yonata Levy. 1999. “Empty Subjects in Finnish and Hebrew.” Natural Language and Linguistic Theory 17:613-71.

11 Responses to “Pro-drop in Hebrew: a summary”

  1. New Blog Underway: Ancient Hebrew Grammar - Berit Hadašah Says:

    […] see that Dr. Robert Homlstedt, over at Ancient Hebrew Grammar, has posted an insightful article on “Pro-drop in Hebrew: a summary.” Dr. Homlstedt has […]

  2. Phil Says:

    Would I be correct in understanding the claim that אכלתי is pro-drop is at least ambiguous, because pronominal information is included in the affix? I would have thought that אני אכלתי would be an example of something like repetition of the pronoun rather than supplying a missing pronoun.

    • robertholmstedt Says:

      That is a good question and the answer entirely depends on what linguistic theory you’re using.

      In generative theory (someone please correct me if a minority position exists within generative), the verbal inflectional (PGN) affixes are *not* viewed as pronouns and thus not the syntactic subject. Rather, they are understood as agreement features signalling grammatical relationships; see, for a good and accessible reference, Crystal 2008* “inflection/inflexion”. This is also the position within some works that have are not explicitly theoretical, such as Waltke and O’Connor 1990 (see p. 691 “inflection”) and Joüon and Muraoka 2006 (see p. 124).

      For those who take the PGN affixes to constitute syntactic pronouns, e.g., GKC 1910:117; Goldenberg 1998**, each case of an inflected verb and overt pronoun represents some sort of dislocation (‘casus pendens’) structure. The problem I see with this, besides the implications for other structures involving pronouns, is that since sometimes the overt pronoun follows the verb, I have difficulty seeing how this would be dealt with syntactically.

      In the generative view, the overt pronoun resides in one of a few syntactic positions that already exist in the phrase structure, so the syntax is fairly straightforward. The presence of the pronoun carries Focus (more often than Topic) features, which also explains why such pronouns typically precede the verb.

      Now it would be nice if a non-generativist could comment on the status of the PGN affixes and pronouns and then also the syntax of a verbal clause with an overt pronoun.

      *Crystal, David. 2008. A Dictionary of Linguistics and Phonetics. 6th ed. Oxford: Blackwell.
      **Goldenberg, Gideon. 1998. On Verbal Structure and the Hebrew Verb (English translation of על תורת הפועל והפועל העברי in מחקרים בלשׁון I, Jerusalem 1985, 295-348). Pp. 148-96 in Studies in Semitic Linguistics: Selected Writings. Jerusalem: Magnes.

      • Ariel G Says:

        Essentially, the structuralist view (advocated for instance by Goldenberg) does not see a principal difference between a morphological marking and a syntactic element. According to Goldenberg, the verb is a complex which contains both the subject and the predicate inside it, and thus also express “nexus” (=predication). In a way, it is the morphological equivalent of a nominal clause which is expressed syntactically by two words (as הנחש ערום).

        The fact that an independent pronoun may appear is immaterial. Wherever it appears (either in the beginning of the clause, in extraposition, or elsewhere) it is not an argument of the verb in the syntactic sense, but rather stands in apposition with the pronoun inherent in the verb. The fact that we have such an appositive pronoun is due to information structure (introducing a new topic, having a contrastive focus, and so forth).

        Although one may argue with the details of this account, I find it much simpler than the pro-drop account, which assumes a) an extra-parameter b) an invisible category c) agreement with an invisible category. In the structuralist view, we have neither of these extra assumptions, and thus should be preferred by Occam’s razor.

      • robertholmstedt Says:

        I am aware of how structuralists like Goldenberg treat the pronoun. And if you consider the approach in isolation, it may win by good Willem’s principle. But since such things cannot be considered in isolation, either in terms of the larger theory or in terms of actually reading the texts from which the data comes, the structuralist notion (at least for Semitics) fails miserably. You are quite wrong, at least according to how Goldenberg describes the redundancy, to say that the appearance of an independent pronoun is immaterial (and this also flies the face of pretty much every other framework). Taking the agreement features to be an actual pronoun necessitates taking the presence of independent pronouns as complex structures (absolutely NOT apposition — the placement of these does not follow the patterns of apposition clear in NWS). It is so much less elegant to follow this backwards linguistics.

        The existence of the structuralist dinosaur in Hebrew studies continues to surprise me. It seems to be a strange Israeli legacy of mid-20th century eastern European linguistics that was ported to Hebrew Univ (and unfortunately brought to a couple US universities). It is tiresome to read and have to deal with.

      • Ariel G Says:

        Well, it seems your approach to this question is biased to begin with, according to your last paragraph. To me it seems that in fact the “old” structuralist approach has gained more ground in recent years, though with new names (be it Basic Linguistic Theory of Dixon, or Construction Grammar of Goldberg). While it is too early to say that the generative approach is on the way out, I believe that in the long run it will be remembered as a scientific anecdote of the 20th century.

        Anyhow, more to the point, I don’t see what are the complication you mention. I am not quite sure I agree with Goldenberg with every detail of his analysis, but in the essentials he is correct: the verbal complex encapsulates a mini-clause, ready with a subject and a predicate, and all the other items of utterance, are floating around this “nucleus”. There is nothing complex about an independent pronoun (or noun phrase) reoccurring beside a pronoun, which happens in Hebrew to be realized morphologically. The same could happen syntactically in Englis : “He, King David, did this”, or in French: “moi, je fais ça”.

      • robertholmstedt Says:

        And here we face the nature of the blogosphere. Kinds of baseless charges that would never be uttered against a professor in a classroom or conference are thrown about willy-nilly. And in this case, the charge is accompanied by great ignorance. Bias implies prejudice and lack of fairness. You know nothing about my linguistics education and so cannot presume to make such a charge apart from ad hominem attack. It’s the same defensive response I’ve faced again and again from the structuralist school of BH grammar. Thankfully, I can say that I had left functionalism and structuralism aside and began investigating generativism for philosophy of language reasons and well before experiencing this type of anti-intellectual discourse, or I may have been pushed away from structuralism out of a real bias.

        Whatever you think about the future of or legacy of linguistic theories is irrelevant. Totally off-topic. And who cares?

        The point here is that the post you reacted to was a commissioned entry on pro-drop in Hebrew for the Encyclopedia of Hebrew Language and Linguistics. I would be biased indeed if in such an article I did not faithfully describe pro-drop from the perspective of the theoretical framework in which it exists. And your English example is backwards, since it is not the pronoun that is appositional, but “King David,” and English doesn’t allow pro-drop anyway. And the French example has no analogue in Hebrew.

        Go read some ancient Hebrew texts, think more carefully about Hebrew grammatical structure, and then come back without charges of bias. You may not be sitting in my classroom, but this is in my blog space.

  3. Vince DeCaen Says:

    This summary on pro-drop is extremely valuable. The bibliography is a gold mine: especially obscure references such as Creason (1991).

    I offer the following comments.

    Missed generalizations?

    (A) para: “Though a full scale of pro-drop … it has long been observed that the subject pronoun is not obligatory …”
    But it is in fact obligatory in important cases, and in the same environments as MH:
    (a) in cases with the participle with/without the tense-bearing aux hyy; and
    (b) verbless clauses.

    (B) para: “One feature that distinguishes …”
    I have difficulty understanding what that paragraph means; but it misses that third person is unmarked among persons, and so can be left unspecified (thus third person is a default), and so need not be a “visible” morphosyntactic feature in MH.

    Contra pro-drop objects: “This is certainly so for Biblical …”

    I would prefer the strong claim that there is no object-pro-drop in BH/MH. In this light, the examples provided are not probative.

    (a) EAT is lexically an activity, and need not take an overt object; in this case, the implicature is partitive. It becomes an accomplishment when an overt object measures out the activity, thus eating the whole thing. Thus in Gen 3: the implicature is just taking a few bites, but not necessarily eating the whole thing (one bite is criminal enough!).

    (b) GIVE has two verbal shells. Thus fruit/apple is a theme sitting in the lower specifier position. The true syntactic object is therefore husband(BH)/me(MH). In this case, there is arguably no object-pro-drop if we insist on “object” as object of some X, and exclude specifier of some XP.

    The research project, then, is to review Creason (1991), and indeed Rizzi (1986), in this light.

    • robertholmstedt Says:

      Thank you for the input.

      Yes, there a number of fascinating issues that I want to investigate, especially the issue of the null object. But since little of that work has been done for BH (apart from the beginnings in Creason’s essays), these were obviously not issues I could take up and defend in the encyclopaedia article (in fact, I was warned against making strong claims that go beyond published studies!). Pro-drop is an issue that must be studied in greater depth for BH. From the nature of your comments, I’m betting we’ll take a slightly different approach (as we did with word order in general).

      With regard to my essay in general, since I have not yet been able to do this with the kind of philological, text-grounded sensitivity that is lacking in so many theoretical studies of BH, I could not in good conscience make any more claims than I did in the encyclopedia entry. Indeed, even this version ran into a few issues with one of the editors.

      I cannot figure out why this paragraph is hard to understand:

      One feature that distinguishes Modern Hebrew is that the null subject may be
      used only with first and second-person past and future verbs; with third-person past
      and future verbs in main clauses and all present tense verbs the subject pronoun is

      Anyone who has studied or speaks Modern Hebrew knows that you cannot have a null subject with 3rd person verbs. It’s a simple asymmetry in Modern Hebrew’s pro-drop paradigm that has been noted by many linguists. And to say that there is a missed generalization about the 3rd person missed an ever larger point — that numerous pro-drop languages do allow pro-drop in the 3rd person. Thus, the asymmetry is worth noting for Modern Hebrew (which is precisely what my paragraph is about).

      • Vince DeCaen Says:

        re bibliography, I would add:

        Panagiotidis, Phoevos. (2002). Pronouns, Clitics and Empty Nouns. [this is a treatment of pronominality in general within a Minimalist approach; a section deals with null subject pro.]

        And I would update Rizzi et sequ., esp. 1997:

        Rizzi, L. (1997). “A Parametric Approach to Comparative Syntax: Properties of the Pronominal System.”

        re paragraph:

        The phrasing is somewhat convoluted, is it not? The young’ns these days can’t follow such syntax.

        Regardless, there is ambiguity (a) with the introduction of “in main clauses” (what then is the status of 3rd person in non-main clauses? and how does this dichotomy apply to 1st/2nd persons?); and (b) the scope of the last “and” is ambiguous (is it just main clauses too? or does the dichotomous qualification fail to apply?).


      • robertholmstedt Says:

        Thank you for the references. I looked at the Panagiotidis book (the library copy was in my office for a good while and now I have my own copy), but I had to be extremely choosey for the biblio in this entry. So, too, regarding the many works of Rizzi.

        The insertion of “in main clauses” is accurate, because subordinate contexts complicate the issue (e.g., null subjects are licensed as resumptives in relatives) and this was not the place for a full discussion.

        As for the ambiguity — none of my readers (nor the editor) had any problems parsing it, and it’s out of my hands now anyway.

Comments are closed.

%d bloggers like this: