Constructive

[kuh n-struhk-tiv] /kənˈstrʌk tɪv/

adjective
1.
or tending to ; helping to improve; promoting further development or advancement (opposed to ):
constructive criticism.
2.
of, relating to, or of the nature of ; .
3.
deduced by inference or interpretation; inferential:
constructive permission.
4.
Law. denoting an act or condition not directly expressed but inferred from other acts or conditions.
/kənˈstrʌktɪv/
adjective
1.
serving to build or improve; positive: constructive criticism
2.
(law) deduced by inference or construction; not expressed but inferred
3.
(law) having a deemed legal effect: constructive notice
4.
another word for structural
adj.

early 15c., “derived by interpretation,” from Middle French constructif or from Medieval Latin constructivus, from Latin construct-, past participle stem of construere “to heap up” (see construction). Meaning “pertaining to construction” is from 1817; “having the quality of constructing” is from 1841. Related: Constructively. Constructive criticism is attested by 1841.

mathematics
A proof that something exists is “constructive” if it provides a method for actually constructing it. Cantor’s proof that the real numbers are uncountable can be thought of as a *non-constructive* proof that irrational numbers exist. (There are easy constructive proofs, too; but there are existence theorems with no known constructive proof).
Obviously, all else being equal, constructive proofs are better than non-constructive proofs. A few mathematicians actually reject *all* non-constructive arguments as invalid; this means, for instance, that the law of the excluded middle (either P or not-P must hold, whatever P is) has to go; this makes proof by contradiction invalid. See intuitionistic logic for more information on this.
Most mathematicians are perfectly happy with non-constructive proofs; however, the constructive approach is popular in theoretical computer science, both because computer scientists are less given to abstraction than mathematicians and because intuitionistic logic turns out to be the right theory for a theoretical treatment of the foundations of computer science.
(1995-04-13)