Introduction The-now-procedures The Harrisian proposal The Harrisian Proposal The aim: compact procedure for describing utterances. Requirements: a list of morphemes and a sequence of morphemes The operation: substitution & repetition Final goal: a general (syntactic) formula applicable to each language yielding a compact statement of sequences of morphemes occurring in that particular language. The rationale Explicitness Extension Syntax The proposal The Proposal Assumption # 1 Complementary variants of a morpheme may be included in a single morpheme. e.g. The set /s, z, əz/ constitute three morphemes yet can be included in one morpheme namely, the plural ending of English -s Assumption # 2 Repeated (identical ) affixation is structured as phrasal infixes. e.g. haiša haktana ( = the small woman) ()האישה הקטנה e.g. Una chica buena e.g. ∂lmullim∂h ∂lз∂meel∂h ()المعلمة الجميلة All these can be regarded not as word affixes or suffixes but rather a single phrase infx The environment To come out with the generalization, we need to set the environment (frame), limitations of occurrence The fewer the limitations, the simpler the syntactic procedure (economy) The operation List Test Class Retest (Subclass) Generalize How it works Morpheme Sequences (144) R, S, T can occur between the frame P ___ Q: (a) Separately: P _R_ Q (b) In a sequence which is taken to be a unit (here enclosed in square brackets): P _[RS]_ Q P [RST] Q In (b), ‘We may say that they are all members of one substitution class [i.e., a unit] which is now not merely a class of morphemes [e.g., ‘N’] but a class of morpheme sequences. Formalizing the System English has these major morpheme classes (among others): T: Determiners; ‘T’ for ‘The’ and the like N: Nouns V: Verbs A: Adjectives D: Adverbs P: Prepositions Hidatsa, a polysynthetic language, has these major morpheme classes (among others): S: Stem P: Prefixes to stems and other prefixes U: Utterance-final suffixes; ‘these occur at the end of (as well as within) within stretches of speech’ (149) F: ‘clause Final suffixes which can be substituted for U if the utterance continues’ (149) N: ‘Non-clause-final suffixes’ (149) ‘What sequences of morphemes can be substituted for single morphemes’? (146) ‘Sequences of morphemes which are found to be substitutable in virtually all environments [frames] for some single morpheme class [e.g., N] will be equated to that class: AN = N’ (146) See page 146 §4.2 for exactly how to determine N1, N2, V1, V2, etc. To limit the generality of these substitutions (so that they are not too general), we use numerical superscripts: N1 + s = N2 Paper + s = papers. Since ‘papers’ is an N2, it can no longer be substituted for N1 (‘paper’). Formalization (cont’d) ‘On the left-hand side of these equations, each raised number will be understood to include all lower numbers (unless otherwise noted) [. . .] On the righthand side, however, each number includes itself alone’ (146). For example: TN2 = N3 is an abbreviation for both: TN1 = N3 and TN2 = N3 The higher the number, the more general the formula. This is because higher numbers can (but do not have to) include more single morphemes than lower numbers!! Affixes (147) Affixes never occur alone. So a V can be turned into an N with an affix, but the affix never appears on its own: ‘-Vn’ means: a suffix (‘-’ before ‘Vn’) attaching to a V (the big ‘V’ is what it attaches to) and creating an N (the little ‘n’ is what it creates). E.g., ‘-er’ in ‘writer’. ‘Nv-’ means: a prefix (‘-’ after ‘Nv’) attaching to an N and deriving a V. E.g., ‘en-’ in ‘enshrine’. An Example from English TN2 = N3: ‘These pointless, completely transparent jokes’ Breakdown of this phrase, without left-hand abbreviations (see slide 16) and rearranged for clarity: T[N2] N2 = [N1 - s] (plural marker; 147 §4.33) N1 = [A2N1] = (147 §4.32) A2N1 = [A2A2N1] (147 §4.32, top right) A2 = [N - Na] (147 §4.33) A2 = [DA1] (147 §4.32, bottom left) D = [A1 - Ad] (147 §4.31, top left) With the colour coding, this sentence is: ‘These [point-less, [complete-ly transparent] joke-s]’ = N3 ☺ ‘The great majority of English utterances’ (149) are at the most general level N4V4. ‘Most utterances of Hidatsa [. . .] consist of S4U’ (150 §5.3). The Theory Itself (1) Position Analysis (and not Morpheme Analysis): ‘strict adherence to morpheme-distribution classes would lead to a relatively large number of different classes’ (150), e.g., think is V, house is N, but take is G (it forms a new class). Variables then are positions, and not morphemes, to limit the number of classes. Thus ‘the element [i.e., morpheme] which occurs in a class position [e.g., V] may be a morpheme which occurs also in various other class positions [e.g., N, D, T, A, etc.]’ (150) (2) ‘The final result for each language [. . .] takes the form of one or more substitution classes [e.g., S4U]. The formulae tell us that these are the sequences which occur’ (150) and that no other sequences may occur which are not ‘derived from the formula’ (150). (3) ‘N has the [positional] values AN, TAN, TA, etc., and [in turn,] each of these [class positions] has specific morphemes as values’ (150) Implicit in the Formulae Suprasegmental Features Morphological Boundaries Morphological [Syntactic] Relations: what affixes actually change the morpheme class; which affixes do not; etc. Endocentric/Exocentric relations (having a head/no head which determines the morpheme class: TN2 = N3/TA2 = N3). Order (the order in which morpheme classes come) (Explicitly) Excluded from the Formulae The following ‘syntactic facts [. . .] cannot be included in [the formulas] and must be found by separate investigations and expressed in separate statements’ (152): Selection Features. These are (mostly) fine-tuned features within each class, creating sub-classes (wirehouse/write-poetry/*‘wire-poetry’ in normal speech). ‘The formulae also cannot in themselves indicate what [semantic] meanings may be associated with the various positions or classes’ (152 §8.3). The formulas do not distinguish what an ‘N’ means from what a ‘V’ means, or subject from object, etc.; it is equally possible to include e.g., ‘◘’ (a completely meaningless symbol) for ‘V’, ‘◊’ for ‘N’. This is why the formulas are syntactic (the way the morphemes/words are strung together into patterns) without reference to their semantics (what the morpheme patterns actually mean).