Pronunciations of 'retard' and 'retardation'
Solution 1:
Because two-syllable nouns tend to acquire first-syllable accents in English, while two-syllable verbs acquire second-syllable accents.
Consider:
- present/present
- desert/desert
- conflict/conflict
- record/record.
See this Wikipedia page about the phenomenon, which includes a list of over 100 words which do this. I remember noticing some words for which this shift seems to be currently in progress, and retard seems to be one of them.
Since the first syllable of the verb is unstressed, the vowel tends to get reduced, and thus changes from /iː/ to /ɪ/ (although some people—me, for one—pronounce the verb with an unreduced vowel). But this vowel change is a secondary consequence of the stress.
Why does this stress shift happen? Once there were enough words that behaved like this, it became a feature of English, which caused more words to undergo this stress shift. I have no idea how it started in the first place, though.
Solution 2:
In reference to people (which is the primary context where I think the pronunciation /ˈriːtɑːd/ is heard) the noun retard seems to be little more than a century old. It's possible that it has affected the pronunciation of the older and more frequent noun retardation, but I actually don't find that especially plausible. The pronunciation of retardation seems to match a broader pattern about the pronunciation of polysyllabic words with multiple stressed syllables.
The verb retard is pronounced with a primary stress on the second syllable, and no stress on the first syllable. Because the first syllable is stressless, the vowel is reduced to /ɪ/ (or in some accents, /ə/). This is a fairly regular stress pattern for a disyllabic Latinate verb where the first syllable is a prefix and the second syllable is a verbal root.
The noun retardation is pronounced with a primary stress on the third syllable (which is regular, because words ending in -tion are pretty much all stressed on the immediately preceding syllable). But that's not the only stressed syllable. Almost all English words have some kind of stress, either primary or secondary, on one of the first two syllables in the word.
There seems to be a certain preference to have stresses in a word fall on alternating rather than adjacent syllables; consider that the first syllable of the verb propose has no stress, but the first syllable of the noun proposition has a secondary stress (and then the second syllable has no stress, and the third syllable has primary stress). See also revolve and repeat vs. the nouns revolution and repetition: the disyllabic verbs have no stress on the first syllable, but the polysyllabic nouns have secondary stress on the first syllable.
In retardation, the situation is complicated because the second syllable doesn't typically undergo vowel reduction, and because of that, it may sound like it has some stress, even though it comes before a stressed syllable. However, it is at least possible to analyze retardation as having no stress on the second syllable: there are some analyses of English stress that allow fully unstressed syllables to have unreduced vowels.
The Oxford English Dictionary marks the pronunciation of retardation as
Brit. /ˌriːtɑːˈdeɪʃn/, U.S. /ˌriˌtɑrˈdeɪʃ(ə)n/
showing stress on the first syllable for both pronunciations, but stress on the second only for the US pronunciation.
Another relevant question is why /riːtɑːdeɪʃ(ə)n/ is used instead of /rɛtɑːdeɪʃ(ə)n/*. When we compare it to words like revolution and repetition, it seems inconsistent, but if we look at a word like elongation that is pronounced with an unreduced vowel in the second syllable, we seem to see a tendency to pronounce prefixes with a "long" vowel in words with this particular stress pattern and syllable structure. Other examples where I think we would often see a "long" vowel in the first syllable: relaxation, resistivity, protestation. As there are no initially stressed disyllabic nouns */ˈriːlæks/ or */ˈriːzɪst/, I don't think it's necessary to suppose that the use of /iː/ in retardation originated directly from the influence of the disyllabic noun /ˈriːtɑːd/: it could have originated from the tendency, whatever it is, that lead to the use of /riː/ in relaxation and resistivity.
(However, that tendency itself might be related to whatever tendency causes /riː/ to be used in disyllabic nouns like retard, redo, refund, and certain pronunciations of resource and research.)
* Walker's Critical Pronouncing Dictionary of 1791 actually does indicate /rɛ/ for the start of retardation, from which I would infer also some level of stress on the first syllable. Unfortunately, Walker did not consistently distinguish reduced vowels from unreduced vowels, so I don't think it's possible to infer from his transcription whether the second syllable could have been fully unstressed or pronounced with a reduced vowel in his time. Another thing to be aware of is that Walker had a prescriptivist bent, and so some of his transcriptions may be based to a greater or lesser extent on rules rather than just on what heard from his contemporaries.