beautiful water of the high place,' _s_ or _[=e]s_ being regarded as
the fractional representative of '_kees_, high.' _Pemigewasset_ has
been translated by 'crooked place of pines' and 'crooked mountain pine
place,'--as if _k[oo]-a_, 'a pine,' or its plural _k[oo]-ash_, could
dispense in composition with its significant base, _k[oo]_, and appear
by a grammatical formative only.

6. No interpretation of a place-name is correct which makes _bad
grammar_ of the original. The apparatus of Indian synthesis was
cumbersome and perhaps inelegant, but it was nicely adjusted to its
work. The grammatical relations of words were never lost sight of. The
several components of a name had their established order, not
dependent upon the will or skill of the composer. When we read modern
advertisements of "cheap gentlemen's traveling bags" or "steel-faced
carpenters' claw hammers," we may construe such phrases with a
latitude which was not permitted to the Algonkins. If 'Connecticut'
means--as some have supposed it to mean--'long deer place,' it denotes
a place where _long deer_ abounded; if 'Piscataqua' was named 'great
deer river,' it was because the deer found _in_ that river were of
remarkable size. 'Coaquanock' or, as Heckewelder wrote it,
'Cuwequenaku,' the site of Philadelphia, may mean 'pine long-place'
but cannot mean 'long pine-place' or 'grove of long pine trees.' If
'Pemigewasset' is compounded of words signifying 'crooked,' 'pines,'
and 'place,' it denotes 'a place of crooked pines,'--not 'crooked
place of pines.'

Again--every Indian name is _complete within itself_. A mere
adjectival or qualificative cannot serve independently, leaving the
real ground-word to be supplied by the hearer. River names must
contain some element which denotes 'river;' names of lakes or ponds