Have the students acquire the new word without the written form, or simultaneously with the presentation of the written form, but with the written form being a backup to many, many exposures to the correct spoken form. Result: no pronunciation problem because of sounding-out errors or over-generalization from the native language, because the Spanish word is already solidly established as its own phonetic form.
The cognate problem is happening because teachers are saying, “See? This is a cognate. It’s easy. It’s just like English.” So students take them at their word. Hmmm….future in English is pronounced “fyu-chure”. Okay, I’ll pronounce it “fyu-tu-ro” in Spanish.
This is why bilinguals don’t have an accent in “the other language”. This is why comprehensible-input based instruction works. Language is not literacy, and cognates are related to literacy, not really to language acquisition. Or only half-related to language, and fully related to literacy. Anyway, it’s easier to “see” a cognate (in alphabetic languages) than it is to hear one, in many cases.
A cognate is ONLY usefully a cognate (for teaching purposes) in the reading sense. Many beginners are not able to recognize Spanish cognates on first hearing unless they are pointed out and maybe even semi-mispronounced to emphasize the relationship to English: “fu-tu-ro, fu-tu-ro, you know, “fyu-tu-ro” like “fyu-ture”? We really don’t want to go there much at all. Having the instructor pronounce “fyu-tu-ro” is just asking for this sort of error to persist. A Spanish student should never heard “fyu-tu-ro” in the context of Spanish being spoken. Which form is the brain going to hold onto — one that’s familiar and similar to the native language, or some more exotic form? Never mind which one is “correct”!
This is related to another of my pet peeves: mnemonics for pronunciation that really do not match the sounds of the target language. They work fine IF the phonetic form of the desired word has been acquired — then they are just a “memory jog” as intended, to prod the learner into recalling a word that he has heard many times but can’t bring up at the moment. Using them as a crutch for words that have not been presented many, many times gives overprominence to the mispronounced elements contained in many mnemonics and allows them to be applied to “figure out” the pronunciation, something that doesn’t come out well in most cases.
This is an excellent argument for making speech and listening primary in the foreign language teaching process, or for delaying literacy even just that little bit until the words have been solidly acquired (not just memorized) through adequate input.
Link the word and the meaning, but give thousands of examples of the word in use, and have the meaning come from comprehensible input (meaningful context). There’s no need to have the phonetic form of the English word come into play at all, unless it’s something done on purpose by the teacher for a specific reason.