A history of the English Language

participle different from the preterite of a verb; as, a book is 
written,
not 
wrote;
the ships 
are 
taken,
not 
took
” With this opinion Dr. Johnson was in sympathy. 
A second consideration was etymology. On this account Johnson and Lowth preferred 
averse from
to 
averse to
. Campbell again states this principle most fully. He says, “When 
etymology plainly points to a signification different from that which the word commonly 
bears, propriety and simplicity both require its dismission. I use the word
 plainly,
because, when the etymology is from an ancient or foreign language, or from obsolete 
roots in our own language, or when it is obscure or doubtful, no regard should be had to 
it. The case is different, when the roots either are, or strongly appear to be, English, are in 
present use, and clearly suggest another meaning. Of this kind is the word 
beholden,
for 
obliged or indebted. It should regularly be the passive participle of the verb to 
behold,
which would convey a sense totally different. Not that I consider the term as equivocal, 
for in the last acceptation it hath long since been disused, having been supplanted by 
beheld
. But the formation of the word is so analogical, as to make it have at least the 
appearance of impropriety, when used in a sense that seems naturally so foreign to it.”
40
By the same reasoning he maintains, “The verb 
to unloose,
should analogically
38 
Philosophy of Rhetoric,
I, 378–79. 
39 
Ibid.,
I, 374–75. 
40 
Ibid.,
I, 397–98. 
The appeal to authority, 1650-1800 265

---

signify 
to tie,
in like manner as 
to untie
signifies 
to loose
. To what purpose is it then, to 
retain a term, without any necessity, in a signification the reverse of that which its 
etymology manifestly suggests?”
41
Fortunately, the third consideration, occasionally made the basis on which questions of 
grammar were decided, the example of the classical languages, and especially of Latin, 
was not so commonly cited. It is true that Johnson is quoted as saying, “It is, seriously, 
my opinion, that every language must be servilely formed after the model of some one of 
the ancient, if we wish to give durability to our works.”
42
Such an attitude derived in part 
from concerns with universal grammar, which Harris defines as “that grammar, which 
without regarding the several idioms of particular languages, only respects those 
principles, that are essential to them all.”
43
Harris was more interested in the 
philosophical problems involving language than in any practical applications that 
discussions of those problems might have.
44
There were other grammarians with more 
normative goals who found it natural to turn descriptive comparisons into prescriptive 
rules, especially since most of the ideas of universal grammar were derived from the 
literary traditions of Latin and Greek. In the course of the eighteenth century a fairly 
definite feeling grew up that there were more disadvantages than advantages in trying to 
fit English into the pattern of Latin grammar, and though its example was called upon by 
one even so late as Noah Webster and is occasionally appealed to even today, this 
approach to grammatical questions was fortunately not often consciously employed. The 
interest in universal grammar for its own sake waned during the following century, and it 
was not until the mid-twentieth century that the works of Wilkins, Harris, and other 
philosophically oriented grammarians in England and France were revived as precursors 
of generative approaches to linguistic analysis (see § 255).