Semantics

Predicates – verbs and adjectives – that take clausal complements differ by the following criterion: The proposition designated by their complement is presupposed by one subset of these predicates, but not presupposed by another subset. A factive predicate is one that presupposes the proposition of its complement; one that does not is non-factive.¹

The matrix verb forget of is factive; the matrix verb contend of is non-factive.

.	a.	The witness forgets that the defendant braked before the collision.
	b.	The witness does not forget that the defendant braked before the collision.

.	a.	The witness contends that the defendant braked before the collision.
	b.	The witness does not contend that the defendant braked before the collision.

The #b versions of both examples illustrate application of the negation test. Both versions of commit the speaker to the proposition that the defendant braked before the collision, but neither version of does.

The set of English factive predicates properly includes the verbs regret, remember, forget, resent, realize; p bothers, matters, as well as the adjectives significant, odd.
The set of English non-factive verbs properly includes the verbs hope, report, assert, think, believe, suppose, imagine; p seems, as well as the adjective true.

Some English verbs allow two syntactic constructions for their complement. For instance, suspect takes a that clause, as in the #a versions of and , but also a Poss-ing construction, as in the #b versions.

.	a.	Mary suspected that John stole the treasure.
	b.	Mary suspected John's stealing the treasure.

.	a.	Mary did not suspect that John stole the treasure.
	b.	Mary did not suspect John's stealing the treasure.

It appears that the #a versions are non-factive, while the #b versions are factive. If so, some English verbs would not by themselves be either factive or non-factive, but instead the entire construction consisting of the matrix predicate and its finite or non-finite complement construction would be factive or non-factive.

¹ The analysis and the terms were introduced in Kiparsky & Kiparsky 1970 and refined in Karttunen 1971.

References

Karttunen, Lauri J. 1971, "Some observations on factivity." Papers in Linguistics 4(1):55-69.

Kiparsky, Paul & Kiparsky, Carol 1970, "Fact." Bierwisch, Manfred & Heidolph, Karl E. (eds.), Progress in linguistics. A collection of papers selected and edited by M.B. & K.E.H. The Hague & Paris: Mouton (Janua Linguarum Series Maior, 43); 143-173.