In informal German, e.g. spoken conversation or text chat, it is possible to omit certain personal pronouns and sometimes inflected forms of sein ‘to be’, too (similar to Russian).

Ich gehe nachhause. → Gehe nachhause.
‘I’m going home.’

Wir sind gleich da. → Sind gleich da. → Gleich da.
‘We’ll arrive there soon.’

Es ist kein Kuchen mehr da. → Ist kein Kuchen mehr da. → Kein Kuchen mehr da.
‘There’s no cake left.’

Es regnet. → Regnet.
‘It’s raining.’

It works best and is most common for first person pronouns, either singular or plural. If a third person pronoun is omitted, it usually implies neuter, i.e. es and neither er (m) nor sie (f). Interestingly, it’s not only used if the inflected verb is unambiguous on its own, e.g. ich bin, du bist, es ist; wir sind, ihr seid, sie sind. However, elision may be suppressed if it would create ambiguity, e.g. not *geh’ nachhause which would be equal to the imperative, although ich geh’ nachhause is pretty much standard now.

Second person pronouns are only omitted in questions, but more often they’re just shortened, becoming a verbal enclitic:

Bist du gleich da? → (Bist d’ gleich da? →) Biste gleich da? → Bist gleich da?

Geht ihr nachhause? → (Geht ’r nachhause? →) Gehta nachhause? → *—

What is the linguistic terminus technicus for this phenomenon? I’m looking for one that is more specific than ellipsis.

  • 4
    The term i'm familiar with is Pro-Drop.
    – P Elliott
    Jul 4, 2014 at 13:14
  • That's one common term, but it presupposes a lot of things about what Pro means. A similar phenomenon in English is called "Conversational Deletion", and it similarly applies to predictable pronouns and auxiliaries and determiners in familiar constructions. Where it might be different is that it only occurs at the beginning of an utterance. Contractions, like Gehta or shouldna, are a different matter, with different etiology and conditions. Of course, they happen simultaneously.
    – jlawler
    Jul 4, 2014 at 18:12
  • @PElliott, Germanic languages are not generally taken to include pro drop. This is an example of a more general ellipsis mechanism that elides material from the left edge of an utterance. Jul 8, 2014 at 11:25
  • 1
    @PElliott: pro-drop means that most pronouns can be left out most of the time in a whole language. This question is about what to call it when pronouns are able to be elided under certain specific conditions only in languages which are not pro-drop. Jul 11, 2014 at 6:31
  • “Pronoun deletion” in Slate’s Lexicon Valley
    – Crissov
    May 8, 2016 at 21:31

2 Answers 2


I have encountered various analyses of and terms for the phenomenon, topic drop being one of them and left edge ellipsis being another. I have used the latter term, i.e. left edge ellipsis (LEE), in my writings.

The examples in the question all submit to an analysis in terms of left edge ellipsis (LEE). LEE occurs in relaxed registers (everyday conversation, emails, text messages, etc.). The ellipsis reaches in from the left edge of the utterance, e.g.

 (Do you want) Coffee or tea?

 (Have you) Been working a lot lately?

 (I am) Quite tired. (I) Will go to bed soon.

LEE cannot occur if the elided material is not at left edge of the utterance, e.g.

 *Now (I am) quite tired.

 *Soon (I) will go to bed.

The German examples in the question all submit to an analysis in terms of LEE. In fact the ellipsis of just a 1st or 2nd person pronoun is a particularly frequent type of LEE in German, English, and many other languages.

I can list a couple of sources (not by me) that discuss the phenomenon in its various forms if anyone is interested.

  • What constrains what can be elided? Of course it's not the case that just any string at the left edge of the utterance can be elided. Do you have a way of doing this using dependency grammar? There are some quite striking asymmetries, e.g. it's possible to say "been working a lot lately?" but not "have been working a lot lately?" - this correlates with the extraction possibilities: "been working a lot lately, have you?", but not: "have been working a lot lately, you?".
    – P Elliott
    Jul 8, 2014 at 12:02
  • @PElliott, I'm not sure about the acceptability judgments. For me, "Have been working a lot lately?" is OK if the elided pronoun is first person: "(I) Have been working a lot lately, (I) Gotta get some rest". On a DG analysis, the elided material at the left edge is a component (but it's often NOT a constituent). See the definition of the component here: en.wikipedia.org/wiki/Catena_%28linguistics%29. Jul 8, 2014 at 12:10
  • @PElliott, the example "Been working a lot lately, have you?" submits to an analysis in terms of topicalization, i.e. the entire VP "Been working a lot lately" is topicalized. Given a DG approach, no such analysis in terms of topicalization is posible for the example "*Have been working a lot lately, you?" because "Have been working a lot lately" is not a constituent (= a complete subtree) on a DG account. Jul 8, 2014 at 12:13

In German, subject pronouns in the Vorfeld (spec-C) can be omitted. (I think Hubert Haider discusses this phenomenon in one of his books.) Note that this is not possible in subordinated clauses:

 1. a.  Ich gehe nach Hause.
    b.  Gehe nach Hause.
 2. a.  Er sagt, ich gehe nach Hause.
    b. *Er sagt, gehe nach Hause.

Note that (2b) is ungrammatical even though a subordinating conjunction is missing in (2.a). The word order in (2a) clearly shows V2-order, which is associated with main clause word order.

Concerning your second set of examples, you are quite correct in assuming cliticization. The position, in which this is possible in many languages, is called the Wackernagel position, technically the second position in the clause.

  • Well, omission only works in indicative. The sub-clause in 2a uses conjunctive (i.e. the final schwa in gehe is mandatory). If it’s direct speech it has unusual punctuation, Er sagt: „Ich gehe nachhause“, and the pronoun could be dropped again.
    – Crissov
    Jul 6, 2014 at 13:53
  • @Crissov I disagree with your entire comment. Your statement that (2a) is in the subjunctive is not accurate. It is a well known fact that subordinating conjunctions can be dropped without incurring the subjunctive. Your second example is misleading. If the subordinated clause were in the second person, it is obvious that subjunctive mood is not necessarily triggered: Er sagt, ihr geht nach Hause. The verb geht is clearly not marked for the subjunctive. Same for the singular. Jul 6, 2014 at 14:27
  • We’re both saying 2b is ungrammatical. It doesn’t matter much (to me) whether that’s because ich gehe is in a subordinated clause (with dropped dass) or because it is in subjunctive mood. How about Ich gehe nachhause, … 1. … weil bin müde, 2. … denn bin müde, 3. … bin müde and even 4. … weil müde bin? All perfectly valid in oral German.
    – Crissov
    Jul 6, 2014 at 15:15
  • @Crissov I'm a native speaker of German, and the instances 1-4 you produced are barely marginal. The point GB theory makes (not that I'm a proponent of theirs, far from it) is that subject pronoun ellipsis can only occur if they are elided from spec-C. Jul 6, 2014 at 18:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.