216 LANGUAGE, VOLUME 74, NUMBER 1 (1998) almost inadvertently, contain information of a sort which cannot now be obtained in the field. [Anthony P. Grant, University of St. Andrews.] Item order in natural languages: Proceedings of LP '94. Ed. by Bohumil Palek. (Acta universitatis carolinae 1994, philologica 3-4.) Praha: Charles University Press, 1995. Pp. x, 415. This collection of papers presented at the 1994 Linguistics and Phonetics Conference held at Charles University in Prague contains 27 papers by 32 authors from seven countries, most papers by scholars from the Czech Republic. AU except one are in English ; the one exception is in German. The linguistic examples used by the authors are in English and Czech for the most part, but German, Japanese, Polish , Spanish, and several other languages are also drawn upon. The articles are grouped according to five topical areas: general considerations (1-52); phonemes, syllables, and suprasegmentals (53-138); word forms in phrases, and phrases (139-244); word forms in sentences (245-377); and text frame (378-415). This is a volume of varied and stimulating contributions as the following sampling of its contents will indicate. That languages are not as neatly patterned and structurally consistent as some tenets of 'classical' linguistics would have them is argued by Frantisek Danes in 'Language is neither all chaos nor all order' (1-9). In 'Templates and melodies: A typology of ordering constraints in early child phonology' (94-124), Kurt Queller puts forth the claim that the speech of a young child is characterized by templates and melodies—two types of sequential patterning that are complementary. A template is defined by the author as 'a word-level package that consists of the entire CV skeleton for a prospective child wordform , in which one or more specific timing (C or V) slots in the sequence may be inherently associated with particular feature values'. A melody is 'a wordlevel package existing on its own autosegmental tier, and consisting of one or more feature values for a specific articulatory parameter' (96). In 'Iconic ordering in morphology' (231-44), Helena Kurzová compares agglutinative and inflecting languages to establish any differences that may exist in the realization of iconicity principles —the iconicity of markedness (quantity) and the iconicity of distance. What typological differences do exist must be taken into account when grammatical structures of languages are compared. Another typologically oriented paper, this one on the sentence level, is the editor's contribution titled 'Typology and item ordering' (357-77). According to Palek, any analysis of different linguistic 'events' in various languages must be conducted consistently by means of a general analytical procedure such as 'the application ofthe predicate ''x is similar to y__" where __ denotes the conditions/circumstances of the event under comparison' (376). In 'Sentence and discourse: An attempt at re-assessment of the background of utterances and sentences ' (383-402), Karel Fíala argues that, contrary to the psycholinguistic approach, there are three levels at which the concept of sentence (utterance ) can be conceived—the syntactic level, the semantic level, and the pragmatic level (with the constituent units of the utterance on the pragmatic level being speech acts). There are occasional lapses in the spelling and syntax ofthe English text, but they never pose any problems in understanding. Given the number and topical variety of the papers, an index at least of subject entries would have been helpful. [Zdenek Salzmann , Northern Arizona University.] Grammar and meaning: Essays in honour of Sir John Lyons. Ed. by F. R. Palmer . Cambridge: Cambridge University Press, 1995. Pp. xiii, 265. This festschrift for Sir John Lyons contains eleven papers—ten by linguists addressing issues Lyons has focused on for 40 years, and one by Lyons commenting on the other papers. The volume also contains a complete bibliography ofthe written works ofLyons. In 'Polysemous relations' (1-25), Adam Kilgarriff and Gerald Gazdar expand on 'regular polysemy ' and argue that a theory of subregularity (less regular but not irregular, i.e. that which approaches homonymy) can be appropriately described using the same formal machinery as other subregular phenomena , such as lexical syntax and inflectional morphology , by using a lexical representation language such as DATR. Adrienne Lehrer and Keith...