24.11.2017 Views

Abelian Groups - László Fuchs [Springer]

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

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

296 9 <strong>Groups</strong> of Extensions and Cotorsion <strong>Groups</strong><br />

Proof. Applying the Hom-Ext exact sequence with Q=Z to e, the exactness of e is<br />

immediate. Its splitting will follow once we can establish that it is torsion-splitting.<br />

Since tC D tC, from the commutative diagram<br />

:0 −−−−→ A −−−−→ B −−−−→ C −−−−→ 0<br />

⏐ ⏐ ⏐<br />

↓ ↓ ↓<br />

• :0−−−−→ A • −−−−→ B • −−−−→ C • −−−−→ 0<br />

it is routine to derive the projective property of torsion groups relative to the exact<br />

sequence e .<br />

ut<br />

The cotorsion hull of a mixed group can be computed easily from those of<br />

the torsion and torsion-free parts. In fact, from Lemma 7.2 it follows that for any<br />

reduced group A,wehave<br />

A Š .tA/ ˚ .A=tA/ :<br />

Example 9.5. Let B D ˚nB n with B n D ˚Z.p n / for each n 2 N. ThenB is equal to the<br />

subgroup G Q n B n such that G=B is the largest torsion-free divisible subgroup in . Q n B n /=B.<br />

Example 9.6. For a reduced torsion-free group A, A Š QA.<br />

Torsion-Free Cover The class of cotorsion groups (and especially their generalizations)<br />

has attracted much attention in recent years, because they form a so-called<br />

cotorsion pair along with the class of torsion-free groups. This means that G is<br />

cotorsion if and only if Ext.A; G/ D 0 for every torsion-free A, andvice versa, A<br />

is torsion-free if and only if Ext.A; G/ D 0 holds for every cotorsion G. Thereis<br />

a very rich, fast growing literature available on this subject for modules. One of<br />

the objects of research is the existence of covers and envelopes (or hulls) in the<br />

respective classes; in our case: torsion-free covers and cotorsion hulls. The latter<br />

has been settled above, so let us move to the question of torsion-free covers.<br />

Needless to say, every group is an epic image of a torsion-free group, for instance,<br />

of a free group. We wonder if there is a minimal one among such torsion-free groups.<br />

Better yet, if there is a unique minimal one. Enochs [2] succeeded in showing that<br />

there is always a minimal one that is even unique up to isomorphism. To make this<br />

fact precise, we have to clarify what “minimality” should mean.<br />

Let W F ! A be a homomorphism of the torsion-free group F into A. is called<br />

a torsion-free cover of A if<br />

(i) for every torsion-free group G and homomorphism W G ! A there is a map<br />

W G ! F such that D , and<br />

(ii) Ker contains no non-zero pure subgroup of F.<br />

Condition (i) says, in more prosaic terms, that maps from torsion-free groups into<br />

A must factor through . Hence it follows that the map has to be surjective (since

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

Saved successfully!

Ooh no, something went wrong!