Factors in the Acquisition of the Present Subjunctive in Spanish : The Role of Reading and Study Jeffery STOKES. Stephen KRASHEN John KARTCH Abstract Universey level students of Spanish were tested on their (acquired) com: ‘petence in the subjunctive. Free reading in Spanish was a significant pre- dictor of subjunctive competence, but length of residence in a Spanish speaking country, formal study, and speifc study ofthe subjunctive were not significom: predictors. These results are consistent with previous research on {froe reading in English a a fist ond second language. Introduction I previous studies, it was reported that actual experience using Spanish was 8 better predictor of competence in the subjunctive than was formal study (Groxs, 1988; StoxeS and KRASHEN, 1990). In this study, we probe deeper {nto this issue in several ways In light of the massive evidence that fee teading contributes to the development of literacy, in fist and second lan- guages (KRASHEN, 1993), a5 well at to the development of competence in complex grammatical structures (LEE, KRASUEN and GRIBBONS, 1996), the ‘mount of fce reading done was added as a posible predictor. In addition, ‘more sensitive measure of formal iatvetion was used. In addition to asking | subjects how many years they had studied Spanish formally, we also asked shout specific study ofthe subjunctive +19.LL Revewof dled Unguses 121-123 (1998) Method Participants were $9 students of Spanish atthe university level, All subjects took teat probing competence in the use of the subjunctive. As in the original study (STOKES, 1988), the test attempted to focus the students on communication, not form Thus, we intended to obtain a sample of their acquired competence, rather than their consciously learned competence Subjects were told tat they were being interviewed as part of a study of attitudes toward college lasswork, and that ony th instructor (..) would hear the tape, Subjects were asked to complete 19 sentences such as the following in Spanish, saying the fis thing tht came to thei mind 1. Siempre sudo cuando 2 Loe pofesorsgeneraimence dan exmenes pars que les etudiates. 3. Bupero conocer a algin prfesor que ‘Test reliability was calculated in two ways, with similar results. Split half reliability was 56 and was .50 according to the Kuder-Richardson formula 20, We discuss the possible consequences of this low level of reliability below. All subjects then filled out a questionnaire asking them to indicate: |. The amount of formal study they had in Spanish (one quarter hour credit of college etady was avazded one point, and one year of high school study was awarded eight points. Ten points were awarded for eight weeks of {intensive missionary training), Length of residence in a Spanist-speaking country (in months); ‘The amount of fee voluntary reading they bad done in Spanish (total number of hous); 4. The adequacy of instruction they perceived they had had on the sub- Jjuetive, rated on a four-point scale (0= non-existent, 1 = superficial, 2= ‘adequate, 3= thorough); ‘Whether they were aware ofthe true nature ofthe subjunctive test. Subjects ‘were asked “Did you at any time during the interview realize the aewal ppurpote of the interview, ie. to gather data on indicatve/ubjunetive use?" Only two students said they realized the purpose of the interview, 20 while even ¢ ‘question. Results and D ‘Table 1 presents foc tree subject on the subjunct (Sroxes, 1988), of Spanish and previous study TABLE ® Ree Parametric stats fiom interval eg cussion, see KEI thatthe predite there is no prob] of formal sy study ofthe subj‘oy hi Fore, At abject ‘Blea. Asin the foeus the stents on 2 sample of their feaaed competence sip of a study of stuctor (1.8) would sentences such es the ‘tei ind glo eins lar results, Splitalf eteRichardson formals awilevelof reliability Wiiges to indicate ‘unre our credit of {of igh hoo sty Editor eit weeks of Binonchs), fein, Spanish (tal Had Bad on the cub f= euperfca, 2 bjadrive test Subjects slew telize the actual fndicativelsbjunctive ‘pose ofthe interview, 1 STORES, RASHEN 5 KARTCHDER while seven said they suspected i, and one subject did not respond to this ‘gestion Results and Discussion ‘Table 1 presents deseriptve statistics. Data on subjunctive study was missing for three subjects, who were assigned a score of 1.3, the overall mean. Scores on the aubjunctive test were eit to those found in the previous study (SroxEs, 1988). Subjeots inthis stady reported somewhat more formal study of Spanish and more foreign residence (mean of foreign sesidence in the previous study was 12.9 months). Deere sais ‘Ament of rma 268 us) Last ofrerierce 293 322 Frereatng 182 esa Sion sdy Toe 6 Parametric statistics were applied to the data, assuming that any violations fiom interval equality were not lage enough to distort the results (for di cession, see KERLINGER, 1976, pp. 402-403). Inspection of Table 2 shows thatthe predictor variables are not highly comelated with each other; thus, there is no problem of ruicollinessity, I is also of interest thatthe amount ‘of formal study of Spanish reported was not correlated with the amount of sudy of the subjunctive Inercontatins ijt ty 10R vetig suey os9 oR 8 oust ring mm oie 1 sijany 20 By O19 195