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.

22 EQUATIONAL LOGIC<br />

some algebras or theories which are one-based:<br />

10.1. The variety <strong>of</strong> all lattices (McKenzie [291]). McKenzie's original pro<strong>of</strong><br />

yields a single equation <strong>of</strong> length about 300,000 with 34 variables. Padmanabhan<br />

[341 ] has reduced it to a length <strong>of</strong> about 300, with 7 variables. Here we mean lattices<br />

formulated as usual with meet and join. Cf. 10.8 and 10.9 below. More generally:<br />

10.2. Any variety which has a polynomial m obeying<br />

m(x,x,y) = m(x,y,x) = m(y,x,x) = x<br />

(a "majority polynomial") and is defined by "absorbtion identities," i.e., equations <strong>of</strong><br />

the form x = p(x,y,...). (McKenzie [291 ]; see also [341 ].)<br />

10.3. Any finitely based variety V <strong>of</strong> EEl-groups (see the beginning <strong>of</strong> õ7)<br />

(Higman and Neumann [ 185] ). Tarski got this for V = all Abelian groups (see [413] ).<br />

(Cf. 10.10 below.) For a recent pro<strong>of</strong>, see [236].<br />

10.4. Certain varieties <strong>of</strong> rings (with operators) (Tarski [413]). For some more<br />

general formulations <strong>of</strong> 10.3 and 10.4, see Tarski [413 ].<br />

10.5. Boolean algebras. (Grgtzer, McKenzie and Tarski) (see [165, page 63]).<br />

(Cf. [401 ].) (Also cf. 10.6 and 10.7.)<br />

10.6. Any two-element binary algebra except (within isomorphism) as in 10.11<br />

below. (Potts [368] .)<br />

10.7. Every finitely based variety with permutable and distributive congruences<br />

(McKenzie [296]; Padmanabhan and Quackenbush [342]). By 9.11 this applies to<br />

any finite algebra which generates a variety with permutable and distributive<br />

congruences, e.g. a quasi-primal algebra (see [362], [369]). Primal algebras were<br />

already known to Gritzer and McKenzie [168]. ([296] contains some very interesting<br />

special one-based varieties.)<br />

Here are some theories (and algebras) which are 2-based but not I-based:<br />

10.8. The variety <strong>of</strong> all lattices given in terms <strong>of</strong> the single quaternary operation<br />

Dxyzw = (xv y) ^ (z vw) (McKenzie [291]). (Cf. 10.1.)<br />

10.9. Any finitely based variety <strong>of</strong> lattices other than the variety <strong>of</strong> all lattices<br />

and the trivial variety defined by x =y (McKenzie [291]). (Here again we mean the<br />

usual lattice operations.)<br />

10.10. Any non-trivial finitely based variety <strong>of</strong> I'2-groups (defined at the

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

Saved successfully!

Ooh no, something went wrong!