Krashen on the size of vocabularies

Rodrigo, V. (2009). Vocabulary size and reading habit in native and non-native speakers of Spanish. Hispania, 92.3, 580-592. [Componente léxico y hábito de lectura en hablantes nativos y no nativos de español]

The results of this paper strongly suggest that non-native speakers who have read a lot have larger vocabularies in the language than native speakers who have not read a lot.

Four groups of subjects took a vocabulary test designed to reflect the size of their vocabulary in the language. The kind of test used is considered standard procedure in vocabulary size estimate studies. It consisted of words chosen at random from a dictionary. If a subject gets 50% of the words right, the assumption is that they know 50% of the words in the dictionary.

The group that did the best were native speakers of Spanish who, it is assumed, had read a great in Spanish (n = 14). All were university students in the US from Spanish-speaking countries.

Second best were non-native speakers of Spanish who had also, it is assumed, read a great deal in Spanish (n = 10). All were graduate students in Spanish language and literature. Only three of the ten were from the US, but none were from Spanish speaking countries.

Third best were native speakers of Spanish who were considered to be non-readers (n = 6). All were employees of the university, and all had immigrated to the US from Spanish-speaking countries in Central America.

The lowest-scoring group consisted of non-native speakers of English who had had little reading experience in Spanish (n = 14). All were students in a fifth semester Spanish class.

1: Spanish L1/readers: mean = 48.6, sd = 5.3. Estimate of vocab size = 48,600.

2: Spanish L2/readers: mean = 40.9, sd = 6.9. Estimate of vocab size = 40,900.

3: Spanish L1/non-readers: mean = 25.5, sd = 9.4 Estimate of vocab size = 25,500.

4: Spanish L2/non-readers: mean = 11.1, sd = 3.3 Estimate of vocab size = 11,100.

Most striking is that group 2 did much better than group 3: The effect size is 2.09, a "huge" effect. Being a reader appears to be more important than being a native speaker.

Rodrigo did not administer a detailed questionnaire on reading habits, nor did she gather data on the precise level of education of subjects in group 3. One could, thus, argue that the results are due to factors other than reading. The results are, however, consistent with the hypothesis that reading was the true cause of the differences, and there are no likely competitors: Studies showing the limitations of formal study make it unlikely that study was the reason for the difference between groups 2 and 3. What is clear is that non-native speakers can reach very high levels of competence in vocabulary, and exceed some native speakers.

Rodrigo points out that it is possible that subjects in group 4 did not show their full competence because of the fact that the distracters were all presented in Spanish on the vocabulary test, and that subjects in group 1 and group 3 who were not from Spain were disadvantaged by the fact that the test used continental Spanish

Subsequent studies can easily deal with these gaps.