Extensive Reading - Limits

Limits

Cobb (2007), McQuillan & Krashen (2008), and Cobb (2008) offer contrasting perspectives. All agree on the need of lexical input, but Cobb (2007; 2008) supported by Parry (1997) convincingly denounces the insufficiency of extensive reading, the current lexical expansion pedagogy, and especialty for confirmed learners. According to Cobb (2007), Krashen (1989)'s influential and seducing hypothesis is that extensive reading generates a continuous hidden learning (lexical input), eventually "doing the entire job" of vocabulary acquisition. This hypothesis is without empirical evidence, neither on the extent (% of global vocabulary acquisition), nor on the sufficiency of extensive reading for lexicon learning (Cobb 2007)

Cobb (2007) thus proposed a computer-based study to quantitatively assess the efficiency of extensive reading. Basically, Cobb estimated the reading quantity of common learners within the L2 language (~175,000 words over 2 years), then randomly took 10 words in each the 1st thousand most frequent words, the 2nd thousand, and the 3rd thousand, to see how many times those words will appear. Those results should be higher than 6 to 10 encounters, the number need for stable initial word learning to occur.

Cobb (2007) summarize as following :" show the extreme unlikelihood of developing an adequate L2 reading lexicon through reading alone, even in highly favorable circumstances" since "for the vast majority of L2 learners, free or wide reading alone is not a sufficient source of vocabulary knowledge for reading". Thereafter, Cobb restated the need of lexical input, and stated the possibility to increase it using computing capabilities.

McQuillan & Krashen (2008) answer that learners may read far more than 175,000 words but rather +1,000,000 words in 2 years. By digging in Krashen & McQuillan own sources and adding some others (Parry 1997), Cobb (2008) convincingly countered Krashen & McQuillan view as being excessively successful cases in reading oversimplified texts. Experiments cited by McQuillan & Krashen use easy and fast to read texts, but not the suitable material to discover new vocabulary. Non simplified texts are far harder, and slower to read. Accordingly, the problem stays at its full strength : common learners need more lexical inputs, extensive reading being insufficient, new sources of lexical input is encouraged to complete it.