I have a question about languages that say "this thing must be the second element in the sentence. For example, German says that in a statement, the finite verb must be the second thing. If you front an adverb, then the subject gets put third. Like this:

ich  ging  nach  Hause
I    went  to    house

"I went home"

heute  ging  ich  nach  Hause
today  went  I    to    house

"Today I went home"

A similar thing happens in Czech. In Czech, a reflexive pronoun (and some other things) have a strong preference to stay in the second position. Czech is also a pro-drop language (this term is apparently out-of-vogue; you are free to suggest a better one). pro-forms are often optional and may be reduced to nothing if their meaning is obvious from something like a verbal conjugation or whatever.

ty   se   osprchuješ
you  REFL shower.PERF.FUT.2PS

"You will have a shower"

osprchuješ           se

"You will have a shower"

How do generative grammars typically deal with this phenomenon? How do they explain that osprchuješ must move up the tree?

  • 1
    I found a chapter from a book that might be relevant: ufal.mff.cuni.cz/~hana/bib/hana-diss-ch4.pdf This chapter itself does not give a generative analysis. Apparently the book has another chapter with an analysis in the framework of Higher Order Grammar. Nov 18, 2016 at 8:12

1 Answer 1


According to a proposal I tried to make here on SE, How can PSG describe the vertical dimension of sentence structure?, height in a constituent structure tree reflects the inherent obliqueness of the various subconstituents. The more oblique a constituent is, the lower it goes in the tree.

Obliqueness is like the traditional notion in declensions that the lower case forms "decline" away from the nominative, except I apply it to phrases rather than words.

Accordingly, constituents move upward in a tree when they become less oblique. This is similar to the Relational Grammar description of English rules like passivization (2 changes to 1) and indirect object shift (3 changes to 2) as "advancements".

Unfortunately I don't see how Relational Grammar, or my version of it, could handle your Czech example. The problem is that verbs other than auxiliaries, do not differ in obliqueness in current versions of the theories.

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.