05.08.2013 Views

Survey 1979: Equational Logic - Department of Mathematics ...

Survey 1979: Equational Logic - Department of Mathematics ...

Survey 1979: Equational Logic - Department of Mathematics ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

20 EQUATIONAL LOGIC<br />

9.18. Some varieties <strong>of</strong> groups (Ol'shanskii [337], Vaughan-Lee [434] )(cf. 13.3<br />

below for a more complete discussion). The existence <strong>of</strong> such varieties <strong>of</strong> groups was<br />

open for a long time.<br />

9.19. The infinite lattice<br />

(McKenzie [ 291 ] ).<br />

9.20. The algebras (2,+,-) (Martin [286, page 210] [287]) and (2,x y) [286,<br />

page 211]. See 8.10 and 8.12 for the definitions and some comparisons. Note also<br />

that (co,+,') is obviously finitely based (cf. 8.1 ). Also cf. õ 12 below.<br />

9.21. The algebra (co,x+y,xy,xY) (Martin [286, page 118] ). (Cf. 8.11.) This is a<br />

negative solution to one version <strong>of</strong> Tarski's "high school identities problem"-he<br />

described a set <strong>of</strong> 8 familiar identities (namely the first 8 <strong>of</strong> Problem 2 <strong>of</strong> õ8), and<br />

asked if these formed an equational base. For another version <strong>of</strong> this problem, see<br />

Problem 2 below and Problem 2 <strong>of</strong> õ 8.<br />

9.22. Any lattice-ordered ring which is an ordered field (and all <strong>of</strong> these have the<br />

same equational theory) (Isbell [ 199] ). (An infinite basis is indirectly described [loc.<br />

cit. ].)<br />

9.23. The variety <strong>of</strong> representable relation algebras (Monk [312] ) and for n 3<br />

that <strong>of</strong> representable cylindric algebras <strong>of</strong> dimension n (Monk [314]). (Roughly<br />

speaking, cylindric algebras are to full logic what Boolean algebras are to logic without<br />

quantifiers"forall,""ther½ exists'.' Relation algebras are intermediate in strength.)<br />

Representable algebras (<strong>of</strong> either type) have a very natural semantic definition; the<br />

definition <strong>of</strong> the entire class <strong>of</strong> cylindric or relation algebras amounts to selecting a<br />

(necessarily rather arbitrary, no matter how utilitarian) finite subset <strong>of</strong> the equational<br />

theory <strong>of</strong> representable algebras. Monk's results indicate that there is really no natural<br />

finite set <strong>of</strong> axioms.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!