Many ESL instructors whom I've met over the past decade taught English abroad before teaching here in San Diego. However, when I started taking classes to get my TESOL Certificate at UCSD, I found that my 'field' experience was quite different from most others'. While I was a foreign language major in high school (4 years of Spanish, two years of French), I had had only three months of travel outside the USA when I went abroad as a graduate student in anthropology.
My real language immersion experience took place while I was in the former Zaire (Democratic Republic of the Congo) in Central Africa studying bonobos/pygmy chimpanzees. This research necessitated language immersion in Lingala while simultaneously making field observations and collecting data on non-human primate subjects. That experience and subsequent returns to Africa greatly influenced my views on adult second language acquisition.
One of my first courses in TESOL was "Theories of Second Language Acquisition." There I was introduced to Krashen's concept of 'affective filter.' What is an affective filter? The concept can perhaps best be understood by an example. When a student is stressed out in the classroom, s/he will put up a protective filter which ironically can also block out the very language s/he's trying to learn. Thus, it is the job of the instructor to provide a safe or sheltered environment in which students can practice the target language, reducing the 'affective filter' effects.
On the other hand, an instructor could consciously ignore the likelihood of anxiety in the foreign language classroom and increase the stress level in order to force production of the target language. What I found in my Lingala immersion in the field is that when I had to depend on the language to communicate my every day needs and wants, my mind, despite the stress, learned relatively quickly how to break the linguistic code. Many people including locals were surprised that, one fine day, I was able to go from cluelessness and sign language to expressing complete thoughts, invent hypothetical situations and argue and defend my need to follow chimps everywhere, even into the swamp forest. For me, it was a lack of consideration on the part of my language 'mentors', rather than a consideration of 'affective filter effects', that pushed me forward into communicativeness in a month's time. I've found in teaching, too, that when students brave the classroom and are put on the spot and say anything that gets a reaction (i.e., makes their classmates nod in understanding, elicits comments or even laughter), these risktakers feel pretty motivated to speak again. If you have a very limited amount of time (i.e., a few weeks) to become productive in the target language, a teacher can create a level of stress through competition.
Games, such as 'Hot Seat,' 'Taboo' with student-made cards, and 'Pictionary', put students under extreme time pressure to speak English so that their team has an opportunity to earn points. Surprisingly, even extremely shy Asian students have blossomed under these pressure circumstances rather than cracked, as some teachers predicted they might.
The bottom line is that if you give people too much time to think about their fear of embarrassment or their shyness, they often give into it. It depends on your objectives as a teacher and the goals of individual students. In the 'real' world, outside the classroom, people need to communicate their wants and needs spontaneously (e.g., in Germany, I went to a pharmacist to get something to stop my daughter's case of the runs, with only one primitive German sentence under control - 'Meine Tochter hat Durchfall', which caused the pharmacist to ask, 'Wie alt?', forcing me to infer from my limited vocabulary - 'Altstadt'- that he was asking how old she was; as I talked to my husband and we realized we didn't know how to say '18' in German, the pharmacist overheard us and shook his head that he understood the number in English; we got the medicine in the correct dose, and 'danke'd the man. This whole episode took less than three minutes, but it definitely raised my confidence in my limited German and proved to me that I could communicate in that language).
Forcing students into a 'spontaneous' situation in the classroom is something that allows them to simulate the real life tension or stress in a safer environment than the 'streets' of a foreign city. My approach to foreign language teaching, consequently, is usually not to lower the 'affective filter' in the classroom, but to use the energy from students' anxiety to 'force' production of English. Accuracy may be sacrificed to some extent, but communicativeness is most students' primary goal.