490 LANGUAGE, VOLUME 58, NUMBER 2 (1982) is a detailed report of the construction, development, and validation of an EFL placement test for German university students. The DELTA is a norm-referenced mix of discrete point and integrative measures, which is shown to be very efficient in predicting later success in learning English, as well as being useful in languageacquisition research. 'Measurement or global assessment of oral foreign language proficiency?', by W. Knibbeler, compares the global and analytic measurement of the French of 108 Dutch adult learners. K attempts to ascertain—via the creation of indices of correctness, fluency, and lexical variety—whether speech acts can or should be described by counting words and errors. The results show a high correlation between the global ratings of proficiency given by trained judges and the ratings derived from quantitative analysis oftranscription ofrecorded speech, the major difference being that global assessment is much less time-consuming. The last three articles are more directly concerned with the quantitative analysis of learner errors. T. Slama-Cazacu and T. Dutescu-Caliban present 'The first hierarchical system of errors (HSE) in the acquisition of English by native speakers ofRomanian', a statistical study from the Romanian-English Contrastive Project of Bucharest. A total of 1183 errors produced by 252 speakers are analysed hierarchically by grammatical structure, taking into account frequency of occurrence and communicative importance. The authors scrutinize the different tasks involved, learner strategies such as 'disguising', the effect of time on learning and errors, and other variables; they conclude that the classical contrastive hypothesis is not valid, since most errors result not from the base language, but rather from influences in the material already acquired in the target language. They outline routes for further contrastive analysis and error therapy. U. O. H. Jung compares the written English of dyslexic and non-dyslexic German school children in 'Dyslexia and the foreign-language teacher'. J first reviews the literature on dyslexia, and then examines the performance of388 subjects on a dictation-cloze test. The results indicate that the classic definition of dyslexia, as characterized by reversal or inversion ofgraphemes in reading and writing, needs to be adjusted (at least in second-language learning situations ), since such a phenomenon represents only 3% of the performance errors; indeed, the 'normal' children tended to reverse and invert letters more often than the dyslexic children. R. Grotjahn closes the volume with a critical assessment of the methodological issues inherent in the quantification of articulatory difficulties —'Zur Quantifizierung der Schwierigkeit des Sprechbewegungsablaufs'. His specific focus is the quantitative definition of the problems which learners encounter in the articulation of German consonant sequences. This volume, which shows substantial awareness of the state of language research in the United States, seems significant for several reasons, not the least of which is the presentation ofthe meticulous research ofEuropean scholars who are not well-known in America. Particularly because ofits critical perspective and dedication to empirical substantiation of learning and teaching theory, the volume will be of considerable value to researchers in language acquisition and proficiency—whether linguists, psychologists, or psychometricians. [Alicia Pousada, Centro de Estudios Puertorriqueños, CUNY.] Deep dyslexia. Ed. by Max Coltheart, Karalyn Patterson, and John C. Marshall. London & Boston: Routledge & Kegan Paul, 1980. Pp. xi, 444. $45.00. In 1966, John Marshall and Freda Newcombe published a case study concerning 'selective dysphasic impairment', with specific emphasis on 'the linguistic aspect of paralexia'. When their patient was asked to read aloud single, unrelated words, he showed an interesting error pattern; instead of saying the given words, he gave semantically-related substitutions, e.g. 'uncle-nephew', 'liberty-freedom', 'little-short' etc. In a subsequent paper (1973) Marshall & Newcombe suggested that these substitution errors indicated the impairment of the spelling -sound route to the lexicon. Within half a decade, the dysphasic profile indicating nonfluent speech, fairly good object-naming, reduced short-term memory span, and predominantly semantic substitutions in oral reading became known as 'deep dyslexia' or 'phonemic dyslexia' (T. Shallice & E. Warrington, 1975). In 1978 a symposium was held at Jesus College, Cambridge, to discuss the 22 cases (16 English and 6 Japanese) reported in the neuropsychological literature. The present book is a result of this meeting. BOOK...