Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Morbocomp German Remarks on the German compound table Anke Lüdeling1 & Peter Adolphs Humboldt-Universität zu Berlin Stems & word classes In the table we use the following word classes and conventions (in compliance with the Morbocomp guidelines). N: Noun stems (in the nominative singular form) and names. Complex names are not internally analyzed. V: Verb stems (unless marked otherwise all verbs are cited as stems); V:inf: Verb infinitives; V:Part1 : present participles ; V:Part2 : past participles. Note that participles often have to be analyzed as adjectives. A: adjectives in the uninflected (predicative) form. Adjectives in comparative or superlative forms are also simply marked as adjectives. Adv: adverbs. Note that in German, adverbs are often not morphologically marked and can be detected only in syntax (many adjectives can function as adverbs). We have therefore given only very few clear examples involving adverbs. P: prepositions Num: numerals. Complex numerals are not internally analyzed. Pro: pronouns Part: verb particles Furthermore we use the following convention: All stems are given in lower case. Note that in German orthography nouns are always capitalized. Linking elements or stem changes In German compounding (and to a lesser extent in derivation) the stem of the non-head element does not always equal the nominative singular stem for nouns or the verb stem for verbs. The possible changes include addition of material (often called linking elements, see below). Examples are hunde+kuchen ‘dog biscuit’ from hund ‘dog’, katzen+futter ‘cat food’ from katze ‘cat’ or lese+lampe ‘reading lamp’ from les ‘(to) read’. elision (mostly concerning schwa syllables). An example is sprach+kurs ‚language class’ from sprache ‚language’. umlauting, alone or in combination with addition of material. Examples are gänse+braten ‘goose roast’ from gans 'goose' and mütter+beratung ‘mothers’ guidance’ from mutter ‘mother’. replacement (used mainly for neoclassical stems or names). Examples are komposita+analyse ‘compound analysis’ from kompositum ‘compound’ or madonnen+kult ‘madonna cult’ from madonna ‘madonna’. Note that the changes do not influence the possibilities for the interpretation of the semantic relationship between head and non-head (there is no 1:1 correspondence between form and meaning). The analysis of this phenomenon is controversial. A linking-element analysis assumes that linking elements are added in the course of the compounding process. This analysis is not easily able to deal with the other phenomena (elision etc.), and it cannot explain which of the possible ‘linking elements’ should be applied to which stem. Another way of analyzing these changes is to assume that the changed form is simply a stem form of the non head element and is learned together with each stem. We cannot give the detailed arguments for either position here, but see Furhop 1998, Eisenberg XXX, Wegener XXX for the background and Lüdeling & Fitschen XXX for an implementation of the stem form analysis.) In the table we give both: the linking element in the LE-1 column. Note that ‘linking elements’ are given purely on a graphemic basis (as in lehrerinnen+konferenz ‘conference of female teachers’ in from lehrerin ‘female teacher’ where we give nen, even though it could be argued that the linking element is only en and the first 'n' appears for independent reasons). Umlaut is marked by a ″. 1 Send any correspondence concerning this table to [email protected]. Morbocomp German in the ANALYSIS column (see below) we give the changed stem and the regular stem in brackets ([Küchen(Küche)<N>+Tisch<N>]<N> for 'kitchen table') The ANALYSIS column We added an ANALYSIS column because some word-internal structures that we wanted to express could not be expressed in the other columns. This column can be taken out for the purpose of the combined table but we think it might be interesting to have if one wants to do more detailed studies. We used the following conventions Complex words are given in square brackets, word classes are marked behind each stem in angle brackets, morpheme boundaries are marked by ‚+’. An example is platzkarte ‘seat reservation’ which is a complex noun consisting of two noun stems and is analyzed as [platz<N>+karte<N>]<N>. Affixes: Affixes that appear within a compound are marked with a ‘-‘. Suffixes are given a word class (following the compounding theory of affixation; see Höhle 1982). An example is [berg<N>+[arbeit<V>+-er<N>]<N>]<N> ‘miner’, where the suffix -er derives arbeiter ‘worker’ from arbeit ‘(to) work’. Stem changes and underlying forms: In those cases where the stem of the non-head word differs from the ‘underlying’ stem (see above), we have given the underlying stem in parentheses behind the stem as it appears in the complex word. Sonnenbrille ‘sun glasses’, for example, which consists of the noun sonne ‘sun’ in the form sonnen and the noun brille ‘glasses’ is analyzed as [sonnen(sonne)<N>+brille<N>]<N>. Similarly, adjective stems in the positive are always given in parentheses when an adjective appears in the comparative or superlative. Likewise, the full form of abbreviations are given in braces, compare [u{untergrund}<N>+boot<N>]<N> ‘submarine’. Depth of analysis: We did not in all cases fully (etymologically) analyze every word. For many deverbal or deadjectival nouns we did not go back to the root because they are fully lexicalised (we are aware that 'lexicalised' is not a well-understood concept). Examples are conversions (without overt marker) like hass ‘(the) hate’ from hass '(to) hate' or derivations with non-productive markers like treue ‘faith’ from treu 'faithful'. Complex names, numerical expressions, prefix verbs or expansions of abbreviations are not internally analyzed. Conversions and transpositions: In order to avoid too many brackets we marked conversions without extra brackets around the converted word and with the original word class in parentheses behind the new word class, e.g. lauf<N(V)> for ‘(the) run’ from lauf ‘(to) run’ or [förder<V>nd<V:Part1>]<A(V:Part1)> ‘helping’. Similarly, for transpositions we mark the original word class in parentheses, compare [einbring<V>-en<V:inf>]<N(V:inf)> for (das) Einbringen ‘(the) positioning’. Problematic cases Zusammenbildungen/verbal compounds/synthetic compounds: the structure of these is not fully clear; they could be analysed as (A+(B+Suff)) or ((A+B)+Suff). Only if we assume the first analysis these should be called compounds. We have decided to list these cases as (A+B), where B is complex. A similar argument could perhaps be given for certain three-element compounds (A+B+C) where neither (A+B) nor (B+C) appear by themselves (Leser 1990, Spencer 1992). Transpositions: We included transpositions (category changes of inflected forms, here: nominalizations of verbal infinitives) in the table. Neoclassical elements/neoclassical compounds: neoclassical compounds are not included in this table. For the purpose of this survey we assume that neoclassical elements like rhythmus ‘rhythm’ or kompositum ‘compound’ are simplex. We were not sure about the columns CLASS, HEAD-S and ENDO. We (kind of) understand the different values for nouns but could not really transfer them to other word classes. For CLASS we tried to decide in all cases but marked some entries by question marks. For HEAD-S we marked some cells with only a question mark. Morbocomp German -mäßig / -los Affixoid References Eisenberg, Peter Fuhrhop, Nanna (1998) Grenzfälle morphologischer Einheiten. Stauffenberg, Tübingen Höhle, Tilmann (1982) Leser, Martin Lüdeling, Anke & Fitschen, Arne Spencer, Andrew Wegener, Heide