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.

WALTER TAYLOR 47<br />

Every finite algebra in HSP A is uniquely a product <strong>of</strong> subalgebras <strong>of</strong> A. Many <strong>of</strong> the<br />

equivalence-invariants <strong>of</strong> these varieties (e.g. the fine spectrum, CEP, AP, Con(V) - see<br />

õ14) are relatively easy to evaluate. See Pixley [362] and Quackenbush [369] for<br />

details and further references - the notion goes back essentially to Pixley, building on<br />

work <strong>of</strong> Foster and Rosenbloom. For infinite analogs <strong>of</strong> primal algebras, see Tulipani<br />

[433] and Iwanik [202]. For congruence-distributivity cf. also 9.11 and 14.7.<br />

Clark and Krauss [89] [90] have given a remarkable theory <strong>of</strong> para primal<br />

algebras, a kind <strong>of</strong> non-distributive generalization <strong>of</strong> quasiprimal algebras, combining<br />

ideas <strong>of</strong> quasiprimality and linear algebra. Also see [371] [300] [173]; cf. 9.13<br />

above.<br />

Pixley and Wille gave an algorithm ([363] [442]; also see [420, Theorem 5.1])<br />

to convert every identity on the congruence lattice (in ^, v, and o) into a Malcev<br />

condition. Which <strong>of</strong> these conditions are "new" remains an open question. For<br />

instance, Nation proved [322] that for certain lattice laws X which do not imply the<br />

modular law, the following holds: if all congruence lattices <strong>of</strong> algebras in a variety V<br />

obey X, then they all obey the modular law.<br />

Very recently S. V. Polin has proved that Nation's result fails for some non-trivial<br />

lattice law X. Non-modular "congruence varieties" are extensively investigated in a<br />

forthcoming paper <strong>of</strong> A. Day and R. Freese. Also see [103], [224] or [298] for<br />

further discussion <strong>of</strong> the state <strong>of</strong> affair just prior to Polin's result. Also see B.<br />

Jdnsson's appendix to the forthcoming new edition <strong>of</strong> [ 163 ].<br />

Another Malcev-definable property <strong>of</strong> varieties V which has received wide<br />

attention is that Fv(n) Fv(m). (See e.g. Marczewski [282], references given there,<br />

and various other articles in the same volume <strong>of</strong> Colloquium Mathematicum.) For<br />

fixed no, the set <strong>of</strong> numbers<br />

{n C co: Fv(n) Fv(n0))<br />

is always an arithmetic progression, and any progression can occur (S'wierczkowski, et<br />

al.). If Fv(n ) - Fv(m ) with m =/= n, then V has no non-trivial finite algebras (Jdnsson<br />

and Tarski [231]) (cf. 14.1 ). (Also see Clark [87] .)<br />

See Csfikiny [95] for a collection <strong>of</strong> properties <strong>of</strong> varieties resembling, but more<br />

general than, Malcev conditions. A nice special example is in Klukovits [247].

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

Saved successfully!

Ooh no, something went wrong!