t, and
arrangement, of words in sentences, constitutes that part of grammar which
we call _Syntax_. But many grammarians, representing this branch of their
subject as consisting of two parts only, "_concord_ and _government_" say
little or nothing of the _relation_ and _arrangement_ of words, except as
these are involved in the others. The four things are essentially different
in their nature, as may be seen by the definitions given above, yet not so
distinct in practice that they can well be made the basis of any perfect
division of the rules of syntax. I have therefore, on this occasion,
preferred the order of the parts of speech; each of which will form a
chapter in the Syntax of this work, as each forms a chapter in the
Etymology.

OBS. 2.--_Agreement_ and _concord_ are one and the same thing. _Relation_
and _agreement_, though different, may yet coincide, and be taken together.
The latter is moreover naturally allied to the former. Seven of the ten
parts of speech are, with a few exceptions, incapable of any agreement; of
these the _relation_ and _use_ must be explained in parsing; and all
_requisite agreement_ between any of the rest, is confined to words that
_relate_ to each other. For one word may _relate_ to an other and not
_agree_ with it; but there is never any _necessary agreement_ between words
that have not a _relation_ one to the other, or a connexion according to
the sense. Any similarity happening between unconnected words, is no
syntactical concord, though it may rank the terms in the same class
etymologically.

OBS. 3.--From these observations it may be seen, that the most important
and most comprehensive principle of English syntax, is the simple
_Relation_ of words, according to the sense. To this head alone, ought to
be referred all the rules of construction by which our articles, our
nominatives, our adjectives, our participles, our adverbs, our
conjunctions, our prepositions, and our interjections, are to be parsed. To
the ordinary syntactical use of any of these, no rules of concord,
government, or position, can at all apply. Yet so defective and erroneous
are the schemes of syntax which are commonly found in our English grammars,
that _no rules_ of simple relation, none by which any of the above-named
parts of speech can be consistently parsed, are in general to be found in
them. If there are any exceptions to this censure, they are very few, and
in treatises still marked with glaring defect