Chris Waigl sends on the cartoon below, displayed in Silver Gulch (“America’s northernmost brewery”) in Fox, Alaska:
For Chris, lager and logger are a minimal pair (with an unrounded vowel in the first syllable of the first, a rounded vowel in the second); for me, they are homophones (with an unrounded vowel in both), which blunts the effect of the joke.
The history and dialectology of low back vowels in English is extraordinarily complex; the Wikipedia entry on the phonological history of the English has a detailed account of the situation, taken from scholarly sources.
With respect to the low back vowels, Chris’s variety of English approximates British RP (“Received Pronunciation”), where there are three phonemically distinct vowels in this phonetic space:
an unrounded long vowel (in father and cart);
a rounded long vowel (in law and caught);
a rounded short vowel (in bother and cot).
My system in this domain is a subtype of GA (“General American”), which has two phonemically distinct vowels:
an unrounded long vowel (in father, bother, and cot; my variety is rhotic, so cart is not directly relevant here);
a rounded long vowel (in law and caught).
Note that I don’t generally have the cot/caught merger that is fairly widespread in American English (usually in favor of an unrounded vowel), but like many GA speakers, I have the merger in some words. As it happens, log is one of them; I have a rounded vowel in dog, but an unrounded vowel in log and also logger (and for some words I have alternative pronunciations), though many GA speakers have a rounded vowel in all three words.
So logger and lager end up being homophones for me (but a minimal pair for Chris Waigl).
(Note: normally I allow comments on this blog, but I’m closing them for this posting, because my experience is that the topic provokes a cascade of unproductive comments about people’s pronunciations of specific words. It’s well known that there are a great many varieties and sub-varieties of English in the domain of low back vowels; that there’s also variation in the treatment of specific words; and that all this variation is associated — but not rigidly — with geography, social class, age, and other non-linguistic factors. Information from particular people about particular words doesn’t advance our knowledge, entertaining though it may be to exchange anecdotes about the way we talk.)