Believe that your child can understand more than he or she can say, and seek, above all, to communicate. To understand and be understood. To keep your minds fixed on the same target. In doing that, you will, without thinking about it, make 100 or maybe 1000 alterations in your speech and action. Do not try to practice them as such. There is no set of rules of how to talk to a child that can even approach what you unconsciously know. If you concentrate on communicating, everything else will follow (p. 26).

These comments are clearly in the spirit of the above discussion.

The position outlined here maintains and refines some of my earlier hypothesis. In Chapter 8 it was claimed that speaking is theoretically unnecessary for acquisition, but may serve to encourage appropriate input (input with the proper size "net"). Acquisition (as opposed to learning), it was claimed, is the result of input, or intake, not actual production. Here, it is suggested that the adult uses the child's output as a part of the information he or she needs to estimate the child's current level of competence. Strictly speaking, however, speaking is not necessary.

We turn now to simple codes the second language acquirer might face, to see whether they might have an effect similar to that of caretaker speech on second language acquisition.

Simple Codes and Second Language Acquisition

All three simple codes discussed above, teacher-talk, interlanguage-talk, and foreigner-talk, are clearly attempts to communicate. The question we thus need to ask is whether the "net" they cast is the same size, and whether they might have the same effect as caretaker speech on language acquisition.

The "facts" about child language acquisition and caretaker speech seem to hold true for simple codes and second languages. In the first place, current research supports the hypothesis that structure emerges in second language performers in a more or less predictable order for adults (Bailey, Madden, and Krashen, 1974; Krashen, Sferlazza, Feldman, and Fathman, 1976; Fuller, 1978; Hyltenstam, 1977), revealed when testing is done in "Monitor-free" conditions (Chapter 4, this volume).

The available data indicate, moreover, that the simple codes we are concerned with are "tuned" to about the same degree as caretaker

128