¹ Methods in structural linguistics 184–6 (Chicago, 1951).

² For the general approach followed here see esp. Noam Chomsky, Syntactic structures ('s-Gravenhage, 1957), and Robert B. Lees, The grammar of English nominalizations (Bloomington, 1960). For a previous application of this approach to German see Lees, Structural Grammars, Mechanical translation 4.5–10 (1957). Part of the work reported here was made possible by a summer research grant from the Excellence Fund of the University of Texas.

³ Chomsky 62.

⁴ A chart of this sort for German is given, for instance, by James W. Marchand in a pedagogical article, The Teaching of German Word Order—a Linguistic Approach, Language learning 8.27–35 (1958). I am, of course, not questioning here the value of such devices for teaching languages.

⁵ Language 168 (New York, 1933). Cf. Sol Saporta's discussion of this point in a review, Lg. 36.93 (1960).

⁶ It should perhaps be emphasized that these symbols have nothing directly to do with German junctural phonemes, although I suspect they will prove useful for the eventual production of such junctural phenomena and the related ones of pitch and stress. One of the main difficulties in the way of writing morphophonemic rules for German is the almost complete lack of good and detailed phonemic studies (in the usual sense) of the suprasegmentals. The slant line is, of course, simply a terminal symbol (with respect to the IC-component) introduced early and repeatedly in the constituent-structure rules.

⁷ My colleague Robert T. Harms has found a similar attack—including the use of a phrase-level boundary marker—useful in dealing with Estonian syntax; see his Estonian grammar (to be published).

⁸ Parentheses enclose sentences with which this study does not deal directly but which are either (6 and 7) part of the relevant preceding context or (10 and 11) examples of problems closely related to those dealt with here. Cf. the discussion of the passive below.

⁹ Structural Grammars, cited in fn. 2

¹⁰ Aside from the special notations explained above, the signs are used in their usual meanings. Separation by commas within braces is equivalent to a vertical listing. The rules given here are minimal, set up mainly to illustrate the treatment of order. In the fuller set of rules many more steps are needed and further distinctions must be made, for instance, between those verbs which can and those which cannot undergo the passive transformation, or between the two types of ein and der for both syntactic and morphological rules. Problems of morphology are not considered, nor the problems presented by the development of Perf as either hab or sei (which must be done after a number of transformations). Bases are spelled in traditional orthography, whereas a fuller treatment will use morphophonemic representations (including one stress morphophoneme).

¹¹ Cf. Karl Boost, Neue Untersuchungen zum Wesen und zur Struktur des deutschen Satzes: Der Satz als Spannungsfeld (Berlin, 1956). Boost calls the ‘topic’ Thema.

¹² Something like this arrangement is implied in the article by G. H. Matthews and Syrell Rogovin, German Sentence Recognition, Mechanical translation 5.114–20 (1958). The shift of the finite verb to final position in a sentence-recognition routine—the inverse of a derivation—is equivalent to a ‘division’ by a transformation like T 3 here.