This is a reply to Mario Saltarelli’s ‘Spanish plural formation: apocope or epenthesis?’ (1970). It is argued that Saltarelli’s one substantive criticism of apocope, his central argument in support of epenthesis, and his defense against anticipated objections are all invalid.

This paper argues that there is a surface structure condition in Spanish which allows only one negative particle no to appear as a constituent of each simplex sentence in surface structure. As a result, the number of grammatical negations has no correlation with the number of S-nodes originally found in deep structure in those trees which have lost S-nodes in the course of their derivation. This hypothesis is supported by structures which have undergone Equi-NP Deletion, by sentences with quantifiers, with adjectives, and with certain adverbs, by certain comparative sentences, and by so-called simple sentences.

Schane 1968 proposes that non-uniqueness of underlying phonological representations be deleted from linguistic description by an application of the markedness principle. The present paper expresses disagreement with Schane’s proposal (§2), adduces Mandaic data interpretable as counter-evidence to that proposal (§3), and discusses alternative solutions (§4). Another type of phonological nonuniqueness is briefly examined in the appendix.

In Maxakalí, some etically disyllabic sequences of the shape [CVCV] are interpreted as emic monosyllables of the shape /CVC/: e.g. [taγӘT] /tat/ ‘to carry’. This paper (a) describes the phonemic system of Maxakalí, with special attention to the unorthodox variants of the consonants; (b) presents the psycholinguistic evidence supporting the analysis; and (c) discusses it in relation to the principle of phonetic similarity.

In this study an attempt is made to show that, upon close examination, certain phonemic alternations in Mandarin Chinese, which hitherto have been regarded as due to dialectal mixing, can be established as being internal evidence for some old phonological distinctions. The study examines various cases of dissimilation among pre-vocalic (or post-vocalic) segments which post-vocalic (or pre-vocalic) segments have conditioned within syllables during the history of the Chinese language. The discussion is focused specifically on the phonological factors conditioning the alternation of velar vs. palatal initial consonants, and of velar vs. dental final consonants. These alternations are interpreted as a special type of dissimilation—that of medials conditioned by final palatal nasal and stop, which can be reconstructed on the basis of quite independent external evidence. This leads to the reconstruction of a four-way distinction among Ancient Chinese ending consonants: labial, dental, palatal, and velar.

This paper presents and motivates a novel ‘epenthesis’ theory of Spanish plural formation as opposed to a recently proposed ‘apocope’ theory. Whereas the apocope theory provides a lexical explanation adding to the complexity of the lexicon and adds a rule, the epenthesis theory provides a phonetic explanation of the phenomenon by generalizing the independently needed rule of prothesis without adding any new rules to the grammar. Simplicity considerations appear to support the epenthesis solution, synchronically as well as diachronically.

Certain examples of English subordinate clauses are introduced in §1, together with the assumptions and evidence used in claiming that a problem, for which the examples are crucial, does in fact exist. §2 contains a brief description of J. R. Ross’s performative sentence hypothesis. §3 shows how Ross’s schema might be modified in order to account for the special problems posed by reason adverbials. §4 is concerned with the derivation of though clauses within a declarative sentence framework. §5 extends the discussion to other types of adverbials.

Constraints on pronoun-antecedent relationships in complex, coordinate, and simplex structures of Samoan are investigated. Three modifications of linguistic theory are proposed to account for the Samoan data. Fillmore’s ‘case grammar’ model is used for the descriptive framework.

This article deals with some semantic categories of shape as coded in grammar. An analysis is given of three patterns in Tarascan grammar: the ‘numeral classifiers’, the covertly ‘classificatory verbs’, and the suffixes of locative space—particularly the semantically complex imagery of the so-called ‘body part suffixes’; the ‘edge-orifice’ suffix is interpreted in considerable detail. The three semantically intersecting patterns are shown to mediate between abstract geometrical features and a concrete map of reference. A brief global survey indicates that many Old and New World languages have similar patterns, and raises questions about the universality of shape categories and their role in abstract thought and metaphor.

Assuming a semantic deep structure that relates to a grammatical surface structure provides a basis for the study of discourse forms in Saramaccan. The common travel narrative, for example, shows a surface organization into paragraphs and sentences with systematic omission of information in some places. To account for this organization we assume a deep structure relation among events, event settings, the narrative setting, the identification of participants in the events, collateral information, and background information.