[Lingo] 黒犬 vs. 黒い犬

["Stephen J. Turnbull" <turnbull.stephen.fw@example.com>]

steve writes:

 > I'm working my way through a Japanes3 3rd grade reader.  It has a short 
 > story about "a small black dog and his small blue coat."  I plugged a 
 > sentence containing 黒犬 into google translate and was surprised to see it 
 > pronounced 「こっけん」as though it was 熟語.  Likewise for 白犬(はっ
 > けん).

My guess is that translate has a rule based on phonology.  (For those
who don't know much technical linguistics, phonetics is stuff like
"English doesn't have the glottal stop 'small tsu', Japanese does, but
it doesn't have 'l'."  Phonology explains how sounds combine.  For
example, in pluralization in English you "add s" in spelling, but in
speech you say pigS, but foxeZ.  It's possible to say foxeS, consider
exceSS, but you don't.)  In Japanese, ○くk- has a phonological rule
that changes it to ○っk- (where "k-" indicates any of かきくけこ).
So here こくけん -> こっけん.

Why does translate apply this rule?  Well, I can't find 黒犬 in the
dictionary (and in my IME くろいぬ -> 黒 + 犬, not 黒犬, of course an
immediate confirmation enters the jukugo).  So it's probably not in
Google's dictionary either.  Without the dictionary entry or the gobi 
い (ie, 黒い犬), translate appears to guess that it's an on-yomi word,
reads the characters as こくけん, and then applies the phonological
rule to get こっけん.

 > I'm guessing this is a problem with written vs. spoken Japanese and
 > that typically one would say 「くろいいぬ」.  And I'm guessing if
 > one were reading the story aloud one would read 黒犬 as くろいいぬ.
 > Is this correct?

Well, Japanese do love their contractions, and here probably 黒犬 is a
descriptive name for the dog, rather than just a description.  I would
imagine that it would be read kun-yomi, as くろいぬ, and 白犬 as
しろいぬ.  If it were supposed to be read くろいいぬ, I would think it
would be written 黒い犬.

Steve