Fact 1

When provided with time to plan their utterances via a writing task, learners produce utterances with greater syntactic complexity and higher lexical density, as measured by the number of words and propositions per utterance (average 21 words per utterance). Yet, errors at the lexical level (word choice) outnumber those at the grammatical level (word form). Grammar consciousness is higher at the writing task. Reciprocal actions of speech encoders would function in favour of accuracy, word form (grammatical encoder) at the expense of fluency (lexical and phonological encoders).   

 Fact 2

When utterances come out freely and spontaneously:

�        Errors are primarily at the level of language appropriateness.

�        Grammatical errors are less frequent because utterances are short, prompt and concise (average 10 words per utterance)

�        Learners display more gambits, lexical dysfluency markers, pauses, hesitation and utterances come out in bribes or chunks.

 Fact 3

When utterances are planned but not preceded by a writing task, grammatical accuracy is lower.

�                      Less use of cohesive devices

�                      Inconsistency in verb tenses

�                      S-V disagreement

�                      �s� inflection (for plural and simple present singular) is frequently missing

�                      Focus is no more on syntactical complexity

Fact 4

In the speaking process, when conceptualization is neither free nor personally motivated but rather triggered by an external stimulus (teacher�s question or student-student interaction) utterance formulation and verbalization would depend primarily on the learner�s lexical repertoire (which have already been provided, acquired or learnt). Reciprocal actions of encoders would function in favour of fluency, word choice (lexical and phonological encoders) at the expense of accuracy (grammatical encoder).    

Fact 5

The longer and more linguistically or syntactically complex the utterance is, the higher the risk of erring would be:

�       Planned utterances score from 21.21 to 39.9 words per utterance with errors from 12% to 11.67% per utterance.

�       Unplanned utterances score from 8.93 to 6.62 words per utterance with errors from 5% to 6% per utterance.