1865:
can't understand. If you put it that way, then surely all of the models which show the independence of various versions of AC and CH are all crippled. So all of model theory is essentially crippled, unless you have some favourite models which you think are the only models. There is nothing misleading in what it written there. It is in the standard literature in peer-reviewed books and journals, like by Tarski, Mostowski, Lindenbaum, Howard, Rubin, Jech, Moore and many others. They all refer to them as definitions of finiteness. I don't think those authors (and many others) are misleading. Just because you can make lots of definitions equivalent by introducing some extra axioms doesn't mean that they are all the same. There's a lot of diversity in set theory. There is not one single consensus set theory (and one model) which is healthy and able-bodied, while all the rest are "crippled". If you take out the diversity, then a hundred years of set theory, proof theory and model theory have been a waste of time.
1247:
without AC of the various finiteness equivalences and non-equivalences. I'm working on some of those right now, like showing equivalence of the
Kuratowski finiteness, the I-finiteness (Tarski) and the enumerative definition. It's so very difficult to see where AC comes in, and so very easy to accidentally use AC unwittingly. The Kuratowski 1920 paper has a very nice short proof of equivalence of his definition and the enumerative definition. It's too bad we can't fill this "finite sets" page with proofs, because I think some of them are very interesting in themselves. I used to think that finite sets were trivial. Not any more! --
931:"In ZF, Kuratowski finite implies Dedekind finite, but not vice-versa. In the parlance of a popular pedagogical formulation, when the axiom of choice fails badly, one may have an infinite family of socks with no way to choose one sock from more than finitely many of the pairs. That would make the set of such socks Dedekind finite: there can be no infinite sequence of socks, because such a sequence would allow a choice of one sock for infinitely many pairs by choosing the first sock in the sequence. However, Kuratowski finiteness would fail for the same set of socks."
1990:
Tarski's definition and the old-style numerical definition of finiteness is that the methods of proving equivalence are themselves interesting methods. I don't know how to express it any better than this. There was a move at a point in history to find definitions for finiteness that did not involve bijections to/from the finite ordinals. Dedekind's and Tarski's and
Kuratowski's definitions arose from that effort. Substituting the old numerical definition for the pure set-theoretic definitions would defeat the whole spirit of the endeavour.
1497:
that model theory most of the time just proves in some way that counterexamples exist without constructing them. If so, then the language "finding counterexamples" would be somewhat precarious relative to the language "demonstrating the existence of counterexamples". On the other hand, maybe the people who don't know about these things don't care, and the people who do know about these things will just skip over it because they know what the situation is.
31:
1643:
finding counterexamples to the pseudo-theorem: "All continuous functions are differentiable almost everywhere." But really, I don't think anyone's going to start trying to build counter-examples within standard ZF on the basis of this wikipedia article. (I have myself tried to find concrete examples of infinite sets which are not
Dedekind-infinite, and I was not surprised that I couldn't find any!)
2213:
is that many authors over many decades sought definitions of finite sets which did not require comparison via bijections to the set ω, whether that class is specified by a set-theoretic predicate or via an axiom of infinity. The purpose of this was to get a "non-comparative" definition of finiteness, not a definition which defines it in terms of equal cardinality to a finite ordinal number.
153:: neither Peano axioms nor the definition of von Neumann representation mention finiteness. Second, finiteness can be easily defined without using natural numbers. Dedekind's definition (having no proper equivalent subset) is mentioned in the article. There is also another definition which does not depend on the axiom of choice (I forgot who invented it, maybe Kuratowski?): a set
271:"Dedekind treats infinitude as the positive notion and finiteness as its negation." What is the point of this sentence? Is it a historical remark? After all, as is later written, "Dedekind finite naturally means that every injective self-map is also surjective": the notion of Dedekind-finiteness is at least as positive as Dedekind infiniteness! OK if I change it?
95:; we're not trying to develop a single well-grounded development of mathematics but instead to describe all of the various interrelationships between mathematical (among other) concepts. So if X can be defined in terms of Y while Y can be defined in terms of X, then both definitions should be included.
2221:
make comparisons to the set ω. Examples of such non-comparative definitions were given by
Sierpinski, Dedekind, Kuratowski, Tarski and numerous others. They certainly knew how to define the class of finite ordinals using a set-theoretic predicate, but they were looking something which did not involve
1943:
page to discover what finite means, and now Ia-finite is being defined in terms of finite and cofinite. That really does defeat the core purpose of the 8 definitions, which is to provide simple self-contained logical predicates using the equality and element-of relations to define finite set concepts
1864:
The first paragraph is clearly true and objective. Yes, CC plus IV-finite implies I-finite. I don't think I can agree with anything in the second paragraph. The subject of different kinds of finiteness is not an axiom of choice matter. Calling the models crippled is an extreme value judgement which I
955:
A "set" lies in a particular model of ZF; some sets are
Dedekind-finite in the model, and some are "Kuratowski-finite". In a sense, a set being "Kuratowski-finite" may be absolute between models of set theory, while a set being "Dedekind-finite" is not. However, that's not relevant to the arguments
2212:
Ignoring any models, it is well known, and stated and proved in numerous mathematical logic textbooks (and in my own book too), that there is a relatively simple closed formula for the finite ordinals, and that this does not require the axiom of infinity. However, that's missing the point. The point
1642:
Frankly I think there was no harm done at all in the changes you made to the wording. As I said before, the experts who know all about it will just ignore it if it's wrong in some small way, and for the non-experts, it has no real effect, except to perhaps create a slightly false idea that it's like
1604:
set-universe to the usual ones. That's why I prefer to be a bit more fuzzy in my language about these things, and talk about proving non-equivalences rather than finding counter-examples. Anyway, that's my 2¢ worth on the matter. I think I've learned enough about finite sets now to last me till Xmas.
1603:
I think my point still remains, though, that the idea of "finding counter-examples" sort of gives a false impression that these are counter-examples in the sense of standard ZF set theory, when in fact the "rules of the game" are quite different, and these counter-examples are in a different sort of
1599:
of atoms has some relation to the set of rational numbers, but the paper presupposes an understanding of
Mostowski models which I do not have. However, the actual choice of set to get the counter-example is not the only thing which is important here. It is the structures which are constructed to get
1496:
Yes, that's true. So I don't exclude the possibility of finding a ZF constructible set which is infinite but not
Dedekind-infinite due to the ZF model being non-AC. However, my point is still that I think this is somewhat rare, although not impossible. I think, unless I see examples to the contrary,
237:
Well, if you're going to phrase it that way you have to be consistent. If the hypothesis is to be stated formalistically (as "what we assume"), then the conclusion needs to be so stated as well, in terms of what we can prove rather than what's true. A bare mathematical statement is equivalent to the
1646:
Getting into model theory in any way, like discussing ZFU or ZFA or
Fraenkel-Mostowski models is really at the edges of the scope of the article. I think it's best to just leave it all simple, with just an intriguing little look at 8 non-equivalent finiteness definitions in ZF, which are equivalent
1282:
And just another little note about this. The Howard/Rubin and Jech definitions of T-finite sets look to me to be absolutely identical to the II-finite definition. Jech gives the T-finite definition on page 52 of his "The axiom of choice" book, which I do possess. And now I have just discovered also
1261:
You were right. The T-finiteness does not belong on that list. It is ambiguously presented in Howard/Rubin. The same definition is given in Jech "Set theory" apparently. (I don't have that book, but the T-finite definition is quoted from Jech on web sites.) It is crystal clear that I-finite implies
2169:
1373:
Now that is fairly bold for 1958 because Paul Cohen's model theory triumphs date from 1963. So they were using a restricted range of models before that time. However. Tarski was an important pioneer in model theory. So I am not surprised that he got some of these reverse non-implications proved in
1848:
This gets to something which has been bothering me about this section — it is really more about the axiom of choice (or the lack thereof) than it is about finite sets. Except for I-finite sets, none of these sets are really finite rather they are victims of the crippled models in which they live.
226:
The reason I changed the line about AC being true to assuming it is that it strikes me as very bizarre to call an independent axiom either true or false. I suppose this is a formalist bias on my part, but suggesting that it actually has some as-yet unknown truth or falseness similarly reflects a
1938:
There are some problems with this. The whole idea is to define finiteness in 8 ways which do not use the traditional definition of ω as a yardstick to measure sets using equipollences or bijections. Now a simple low-level set-theoretic definition has been replaced with the concept "finite" which
1246:
Actually, I don't know how to do most of the proofs. I just trust what Howard and Rubin wrote in their really excellent "Consequences of the axiom of choice" book. It has such good binding, I figure it must be true! But perhaps I could just comment that I find it fascinating to find proofs in ZF
1985:
Yes, however, Zorn's lemma is equivalent to the axiom of choice, but we don't say that they are the same thing. In fact, I-finite and finite are only equivalent in full ZF. I think they are probably not so equivalent without the axiom of infinity. How do you measure sets against finite ordinal
1650:
What I've read of the Lévy article is pretty much the discovery of counter-examples. So I'm happy to leave the wording as you have put it. But personally, I prefer not to get into model theory. I think it's all a bit of hocus-pocus and rabbits out of a hat. Model theorists use ZF and other set
1989:
But the real issue is not what is equivalent to what. The real issue is what is the definition. A definition is a particular logical expression, which may or may not be equivalent to other definitions under particular ranges of circumstances. The whole point of having
Kuratowski's definition,
1969:
is either I-finite or co-I-finite or both.". But I thought that would sacrifice some of the benefit of a simpler statement. And the line just above says that I-finite "is also equivalent to the normal definition of finiteness". That is, finite = I-finite. This is also stated in the section on
1423:
My understanding of this subject is not at all expert, but from the little that I do know, I think that in model theory one rarely actually constructs concrete counterexamples. For example, I would dearly love to see just one set which is infinite, but not
Dedekind-infinite. Or a set which is
1947:
Of course, the definition of Ia-finite in terms of finite and cofinite is simpler when you know and understand ω, the axiom of infinity, recursion, and so forth. But that is defining low-level concepts in terms of high-level concepts. I think that Ia-finite should be defined in terms of
1427:
It seems to me that it is possibly misleading to say that one "finds counter-examples" in models for various non-theorems. Wouldn't it be more correct to say that one "demonstrates the existence of counter-examples"? That would be a better form of language if the counter-examples are
939:
the case that "there can be no infinite sequence of socks" must be true when AC is not assumed. This should imply that such a set can not be determined in ZF to be either Dedekind-infinite or Dedekind-finite! This is very different from asserting it is Dedekind-finite.
241:
It could be phrased something like "the following are provably equivalent in a formal theory including the axiom of choice". But I think that's just awkward; the usual convention is to state things Platonistically, and let formalists reinterpret as they will.
699:
The article on Dedekind-infinite says "... that a set is infinite if and only if it is Dedekind-infinite. ... The full strength of AC is not needed to prove the equivalence; in fact, the equivalence of the two definitions is strictly weaker than the
1844:
Countable choice and IV-finite (Dedekind finite) imply I-finite. So countable choice is enough to show that I, Ia, II, III, and IV are the same. After that, I think a stronger form of AxCh is needed to establish equivalence of those with V, VI, and
1647:
in ZF+AC. (I think the equivalences just need countable choice, but I can't swear to the truth of that.) After all, wikipedia is only an encyclopedia, not an academic journal. Readers have to read the references if they want to get real knowledge!
1590:
of atoms of a Mostowki ZFA model as the basis for all counter-examples. So in this sense, the basic example set is in fact a normal set (in fact countably infinite) within ZF, and not a normal set in a model where AC is false. (This is related to
2216:
In other words, the issue is not the length of the logical predicate which defines finite sets, nor whether an axiom of infinity is required a-priori. The issue is whether one can plug any set into a simple set-theoretic predicate which does
1283:
that Howard Rubin on page 184 say explicitly that T-finite and II-finite are the same. That confirms it. So I don't know why they didn't say so on pages 278–280. Now I'm beginning to think you can't judge a mathematics book by its binding!
1262:
T-finite when the definition is written correctly. But I don't know how it fits into the rest of the scale. I think the remaining 8 definitions are solid though. By the way, if you google "T-finite" right now, it points to the
1124:
I believe I understand this now; thanks much. (Although model theory isn't needed, my lack of foundation with it is one of the reasons I originally misinterpreted this section. This is why I'm currently reviewing Enderton's
2052:
1369:"In Tarski (1954) and Levy (1958) it is shown that if a set is finite according to any definition in the above list then it is finite according to any following definition and none of the reverse implications hold."
1993:
What I have done several minutes ago is to simply remove the double-negatives so that it read better. Otherwise it is the same as the original, since the equivalence of not-not-P to P is a purely logical
1627:
What bothered me about your wording was that when you talked about proof, you did not make it clear that you were talking about proving the non-provability rather than proving the reverse implication.
1428:
non-constructible. But I would be happy (and grateful) if someone could give me a constructible infinite set which is not Dedekind-infinite. Then the Bolzano-Weierstraß theorem would bite the dust.
1048:
Hopefully you all can parse what I'm saying--I'm very, very rusty on model theory and tend to use all the wrong words. I just wanted to double-check that my intuition was not the case here.
1696:
exists. Sometimes these choices are wrong and we must back-track and take the other way. But eventually (if there is not a contradiction), we will get to a model where all the properties of
1621:
Some of the independence results may be relative to ZF with urelements (ZFU) or with atoms (no regularity) as you say. So perhaps we should change the article to refer to ZFU rather than ZF.
2039:
Even if the natural numbers do not form a set (i.e. ω does not exist), one can still talk about the natural numbers because there is a predicate which separates them from other sets (see
892:
Arthur, if you (or anyone) has doubts about any part of these proofs, I will be happy to provide more details to convince you. Just indicate which part you want explained in more detail.
2014:
I appreciate your desire to avoid referring (indirectly) to natural numbers. However, I have shown elsewhere that the natural numbers can be defined as a proper class, if necessary.
1166:
to the article. Among other things, it says that I-finite (i.e. finite in the normal sense) does not imply T-finite in ZF without AxCh. I find this difficult to believe. If a set
1651:
theories to construct models to make statements about ZF and other set theories. So it's all a bit circular, especially since the absolute consistency of ZF can't be proved.
935:
Isn't this technically incorrect? AC is not decidable in ZF, so there are models in which such a sequence exists, and others in which it does not. So it is expressly
134:
Well, there exists no bijection between the empty set and one of its proper subsets, since there are no proper subsets of the empty set, so by default it is finite.
1668:
To show that the reverse implication (from III-finiteness to II-finiteness for example) is not provable, one needs to show that there is no deduction from ZF plus "
1730:
That is more complicated than necessary (or than commonly used by most set theorists). The standard approach is to assume you have a countable standard model
112:
I don't agree : even without being Bourbaki, if one definition is given in terms of the other and reciprocally, then it is mathematically completely useless !
667:
OK, I should have thought of that. It's not (almost) immediately obvious, like the other equivalences, though. I can't find my copy of Rubin and Howard,
2164:{\displaystyle \forall M(\forall r\in M\forall s\in r(s\in M)\Rightarrow \forall n\in M(n\in \omega \Leftrightarrow M\models \operatorname {NatNum} (n)))}
1886:
is equivalent to the full axiom of choice. So VI and VII need the full axiom of choice to be equivalent to I. So the only one which is unclear is V.
2238:
1159:
889:. So by induction, this stronger property of partial orderings holds for all finite sets. So the last condition in the equivalence also holds. QED.
1939:
invokes existence of bijections to and from ω, and cofinite, which most readers know even less well than "finite", and presumably they came to the
1424:
Dedekind-finite but not finite. A concrete set of this kind would effectively prove that countable choice is false, not just independent of ZF.
257:
The first occurrence of the term 'Zermelo–Fraenkel set theory' was a cryptic unlinked acronym: ZF. I added the full term and wikilink'ed it. --
1300:
To Alan: Thanks for fixing the subsection and providing references. In particular, thanks for removing T-finite as a kind of super-finiteness.
1303:
However, the subsection still seems unclear on whether some of the reverse implications work — saying no in one place and yes in another.
2017:
Some people are familiar with the filter of cofinite subsets or the ideal of finite subsets. I wanted to suggest that the powerset of
407:
1522:. That is, some of their subsets or functions involving their elements or subsets and functions of their powersets are missing from
721:
The condition about partial orderings is equivalent to saying the (seemingly stronger) statement that for every partial ordering of
197:, but lots of the links are not in fact about finite sets, but about other notions. There should really be a disambiguation page at
1624:
Certainly some aspect of proof is involved — one must prove that a counter-example exists and that it is in fact a counter-example.
1685:
1883:
1170:
is finite, then a simple application of mathematical induction shows that its powerset is also finite since the powerset of
118:
In fact, I wanted to know if according to wikipedia, a finite set can be empty (i.e., if the empty set is finite or not).
311:
2234:
2001:
1955:
1872:
1658:
1611:
1600:
a proof of the non-existence of equivalences between the various definitions. Those might be somewhat non-constructive.
1504:
1435:
1409:
1385:
1290:
1273:
1252:
1153:
38:
1704:, then we can get its rank and determine what its elements are and their elements, etc.. Thus the counter-example
364:
There is redundancy between the Basic Properties and Closure Properties sections. Perhaps they should be merged?
701:
672:
411:
1688:, one can do this by constructing a model. That is, one makes a series of choices about what the properties of
1069:
I do not think that model theory is necessary or helpful to interpret the paragraph which you quoted. Suppose
683:(which should both be there), I'd consider it proved, although I shouldn't add it as a reference, myself. —
2230:
2184:
1997:
1951:
1868:
1654:
1607:
1555:
1500:
1431:
1405:
1381:
1286:
1269:
1248:
1149:
1134:
1053:
945:
261:
298:
227:
strong realism, while simply making mention of an assumption gives a more neutral perspective, I think. --
1832:
980:
showed, this can be cut-down to a model of ZFC (in which there are no such sets). It is not the case that
960:
687:
401:
1568:
1130:
1049:
941:
2226:
369:
325:
102:
mention the other property should also be included! If you have any good ones, then please add them.
2200:
2192:
2026:
1975:
1891:
1854:
1773:
1761:. I recall reading that if Q is a finite subset of the axioms of ZF, then ZF ⊢ Consis(Q), so that
1721:
1632:
1531:
1486:
1355:
1236:
1211:
1114:
1027:
993:
972:
Suppose we are given a model of ZF containing a Dedekind-finite (Kuratowski-)infinite set, call it
897:
709:
657:
1222:
OK. Now I see that the axiom of choice may be needed to select the order in which the elements of
745:, this is easily proved by induction. For the empty set, it is vacuously true because there is no
1002:
That's true. However, the model might be contained (indeed, because AC is Platonistically true,
478:
1592:
2188:
2040:
1829:
957:
684:
397:
106:
347:
293:, in this article. This interesting section introduces infinite set theory, but the article
286:
228:
2222:
von-Neumann-style ordinals at all. (Nor any other kind of ordinal number representation.)
1581:
1398:. This is the right paper to look at! It's got all of the 8 definitions that I put in the
1082:
912:
676:
450:
383:
365:
272:
135:
47:
17:
1010:
is present, but is no longer Dedekind-finite. In the larger model, there is a map from
2196:
2022:
1971:
1887:
1850:
1717:
1628:
1543:
1527:
1482:
1351:
1232:
1226:
are added to the empty set and thus to determine the bijection between the powerset of
1207:
1110:
1023:
989:
893:
705:
653:
243:
212:
202:
142:
78:
1547:
1395:
1595:, of course, where a set can be both countable and not countable, etc. etc.) The set
977:
442:
351:
128:
1738:
of ZF (or ZFU) with the desired characteristics. If the desired characteristic is "
1676:
is II-finite". This is the same as showing that there is no deduction from ZF plus "
331:
258:
1105:
consisting of pairs of socks which have a nonempty intersection with the image of
1089:
is finite. Obviously this contradicts the axiom of choice. It also implies that ∪
2011:
Yes, I think that the double negative was a big problem. I am glad you fixed it.
318:
46:
If you wish to start a new discussion or revive an old one, please do so on the
2242:
2204:
2030:
2005:
1979:
1959:
1895:
1876:
1858:
1835:
1765:⊢ finite fragments of ZF + (desired propery) correspond to finite fragments of
1725:
1662:
1636:
1615:
1535:
1508:
1490:
1439:
1413:
1389:
1359:
1317:
I wonder whether some of the definitions might be equivalent to the following:
1294:
1277:
1256:
1240:
1215:
1138:
1118:
1057:
1031:
997:
963:
949:
916:
901:
713:
690:
661:
386:
373:
354:
275:
246:
231:
215:
205:
176:
1940:
1399:
1263:
1163:
908:
680:
380:
294:
290:
194:
1934:
is not equal to the union of two disjoint sets neither of which is I-finite."
1469:
are not necessarily the same as the properties which it has as an element of
1402:
wiki page. The reverse implications are all disproved using Mostowski models.
1022:
is Dedekind-finite, because the original model does not contain this map. --
1692:
are, and also what the properties are of such other sets as must exist if
1374:
1954. This article has some info on Levy's contributions to this problem:
1006:
contained, assuming the model is wellfounded) in a larger model in which
124:
121:
92:
168:: empty set is finite, according to everybody, not just Knowledge. Yes,
1375:
1186:) is equivalent to the disjoint union of two copies of the powerset of
464:
To show that they are all equivalent, it is sufficient to show that if
304:
There are at least four possible places where this paragraph could be:
173:
115:
Also, if n=0 is a natural number, then is {1,2,3,...,n} the empty set?
2041:
Axiom of infinity#Extracting the natural numbers from the infinite set
1101:
would allow one to define a choice function on the infinite subset of
198:
190:
182:
150:
82:
1314:, making that similar to the next two definitions in the subsection.
1849:
Calling them alternative definitions of finiteness is misleading.
1772:
For those confused by metamathematical paradoxes, I know that the
253:
minor edit: add linkage to the term 'Zermelo–Fraenkel set theory'
2021:
is merely the union of that filter and that ideal in this case.
1769:⊢ ZF, so for any finite fragment of ZF + (desired property), ZF
1477:
to some proposition while still satisfying that proposition in
201:
that would point also to other notions of finite quantities. --
1145:
Is it really true that I-finite does not imply T-finite in ZF?
628:
is surjective but not injective. Let the partial ordering put
25:
1986:
numbers if the set of finite ordinal numbers does not exist.
1518:
they will lack some of the superstructure which they have in
753:. Let us make the inductive assumption that it is true for
189:
Adjectives make bad article titles. I've moved what was at
1419:
Does one really "find counterexamples" in ZF model theory?
1202:) which is the same as the definition of T-finiteness of
489:. Using mathematical induction one can show that all the
1365:
The Howard/Rubin book is fairly unambiguous about this.
172:
is a sloppy notation for the empty set in this case. --
1712:. However, proving that may require assumptions (about
1548:"The independence of various definitions of finiteness"
1514:
The counter-examples will all be infinite sets. But in
468:
is infinite, then the other three conditions fail. Let
457:
Do you agree that it is obvious that the finiteness of
98:
That said, any interesting definitions of X or Y that
2055:
1396:
The independence of various definitions of finiteness
1970:
Necessary and sufficient conditions for finiteness.
1073:is an infinite set of pairs of socks and for every
346:You are invited to give your opinion about this in
2163:
1306:I notice that Dedekind finite is equivalent to 1+|
310:in a new separate article, such as, for instance,
1684:is not II-finite" to a contradiction. Following
1394:The Levy (1958) paper is available online here:
984:changes its character in this submodel. Rather
414:is sufficient to establish the equivalence of:
157:is finite iff every nonempty set of subsets of
511:∈ω} is a countable set of non-empty sets. Let
281:Moving section "Foundational issues" elsewhere
1018:. The original model can delude itself that
652:, then this ordering has no maximal element.
8:
1700:are specified. If we know the properties of
1473:. So it could serve as a counter-example in
285:After weeks of navigation in Knowledge (see
211:Done (but it should have lots more links) --
2179:using ZF, even though the transitive model
1481:. That is, properties are model dependent.
560:which implies the Dedekind-infiniteness of
1164:Finite set#Other definitions of finiteness
252:
161:contains a maximal element wrt inclusion.
2054:
1093:is Dedekind-finite because any injection
596:+1) is injective but not surjective. Let
1162:) just added the interesting subsection
1965:I thought about saying "Each subset of
1918:is either finite or cofinite or both.",
1541:The 1956 paper by Lévy on finite sets (
1198:) which would imply II-finiteness of P(
572:to itself unless it is in the range of
523:) be the first element in the image of
1882:VI→I is equivalent to ¬I→¬VI which by
1577:
1566:
907:Do you intend to add source material?
424:(Dedekind-finite) Every function from
44:Do not edit the contents of this page.
781:be the element which is maximal over
299:Basic concepts in infinite set theory
297:is not even included in the category
77:This encyclopedia defines the terms "
7:
829:which contradicts the maximality of
461:implies the other three conditions?
1734:of ZF + V=L, and construct a model
669:Consequences of the Axiom of Choice
334:(see suggestion by JRSpriggs below)
312:Introduction to infinite set theory
238:statement being true, not provable.
2107:
2077:
2065:
2056:
24:
923:Error in Dedekind-finite example?
410:) questioned whether ZF plus the
988:simply is not in this submodel.
29:
1750:is a new constant symbol, then
428:one-to-one into itself is onto.
2158:
2155:
2152:
2146:
2131:
2119:
2104:
2101:
2089:
2062:
1194:would imply I-finiteness of P(
902:23:53, 29 September 2009 (UTC)
714:00:43, 26 September 2009 (UTC)
691:15:27, 25 September 2009 (UTC)
662:11:35, 25 September 2009 (UTC)
1:
1884:Tarski's theorem about choice
1139:19:43, 1 September 2011 (UTC)
1109:. That is what it is saying.
737:which is greater or equal to
733:, there is a maximal element
1686:Gödel's completeness theorem
1127:Mathematical Intro. to Logic
675:is listed as implying every
536:) which is not equal to any
515:be its choice function. Let
289:), I have found the section
216:22:15, 26 October 2005 (UTC)
206:20:41, 25 October 2005 (UTC)
2243:07:21, 15 August 2014 (UTC)
2205:07:03, 15 August 2014 (UTC)
2031:10:46, 11 August 2014 (UTC)
2006:09:52, 11 August 2014 (UTC)
1980:09:42, 11 August 2014 (UTC)
1960:07:40, 11 August 2014 (UTC)
1896:17:14, 12 August 2014 (UTC)
1877:12:21, 12 August 2014 (UTC)
1859:12:08, 12 August 2014 (UTC)
1836:04:12, 11 August 2014 (UTC)
1726:16:18, 10 August 2014 (UTC)
1663:10:52, 10 August 2014 (UTC)
1637:10:35, 10 August 2014 (UTC)
1616:10:12, 10 August 2014 (UTC)
1536:08:42, 10 August 2014 (UTC)
1230:and its finite cardinality.
1119:06:27, 31 August 2011 (UTC)
1058:03:47, 31 August 2011 (UTC)
1032:09:14, 22 August 2011 (UTC)
998:02:31, 22 August 2011 (UTC)
964:00:06, 22 August 2011 (UTC)
950:23:39, 21 August 2011 (UTC)
568:which maps each element of
355:10:07, 31 August 2007 (UTC)
177:18:28, 14 August 2005 (UTC)
85:" in terms of each other!!
2260:
1922:whereas before there was:
1509:14:02, 9 August 2014 (UTC)
1491:13:56, 9 August 2014 (UTC)
1449:is a model different from
1440:12:17, 9 August 2014 (UTC)
1414:09:24, 9 August 2014 (UTC)
1390:09:01, 9 August 2014 (UTC)
1360:07:43, 9 August 2014 (UTC)
1295:03:41, 9 August 2014 (UTC)
1278:03:26, 9 August 2014 (UTC)
1257:08:24, 8 August 2014 (UTC)
1241:08:01, 8 August 2014 (UTC)
1216:07:54, 8 August 2014 (UTC)
1206:. What am I missing here?
849:is maximal over itself in
741:. For (Kuratowski) finite
556:is an injection from ω to
435:onto itself is one-to-one.
393:Countable choice is enough
387:10:34, 27 April 2009 (UTC)
374:04:26, 27 April 2009 (UTC)
344:Please do not answer here.
276:10:52, 1 August 2007 (UTC)
1746:is not II-finite", where
1708:should actually exist in
865:in which case there is a
702:axiom of countable choice
673:axiom of countable choice
412:axiom of countable choice
262:22:16, 13 July 2007 (UTC)
131:23:29, 19 Apr 2005 (UTC)
109:06:46, 25 Sep 2003 (UTC)
2047:is that predicate, then
837:. On the other hand, if
247:20:36, 12 May 2006 (UTC)
232:18:52, 12 May 2006 (UTC)
2185:axiom of extensionality
1757:is a counterexample in
1556:Fundamenta Mathematicae
1190:. Thus I-finiteness of
917:18:05, 6 May 2011 (UTC)
2165:
1799:finite → (ZF ⊢ Consis(
1716:) far beyond just ZF.
1461:, the properties that
1014:to a proper subset of
869:which is maximal over
2166:
1788:)) ↔ Consis(ZF), but
1465:has as an element of
644:(2)<... for every
337:just where it is now.
42:of past discussions.
2191:(which includes the
2053:
793:is still maximal in
725:, for every element
648:not in the range of
431:Every function from
326:Axiomatic set theory
185:should be a dab page
149:defined in terms of
70:Circular definition?
2193:axiom of regularity
1808:is not the same as
1774:compactness theorem
1672:is III-finite" to "
1376:Levy and set theory
1345:is Dedekind finite.
1331:is Dedekind finite.
809:must be maximal in
620:be the identity on
498:are non-empty. So {
479:injective functions
291:Foundational issues
2231:Alan U. Kennington
2189:axiom of induction
2183:may only obey the
2161:
1998:Alan U. Kennington
1952:Alan U. Kennington
1944:of various kinds.
1869:Alan U. Kennington
1742:is III-finite and
1680:is III-finite and
1655:Alan U. Kennington
1608:Alan U. Kennington
1501:Alan U. Kennington
1432:Alan U. Kennington
1406:Alan U. Kennington
1382:Alan U. Kennington
1287:Alan U. Kennington
1270:Alan U. Kennington
1249:Alan U. Kennington
1150:Alan U. Kennington
441:is empty or every
2246:
2229:comment added by
2175:can be proven in
1914:. Each subset of
1906:So now we have:
67:
66:
54:
53:
48:current talk page
2251:
2245:
2223:
2170:
2168:
2167:
2162:
1819:finite → Consis(
1784:finite → Consis(
1593:Skolem's paradox
1585:
1579:
1574:
1572:
1564:
1552:
927:With regard to:
881:is maximal over
624:otherwise, then
443:partial ordering
421:is a finite set.
287:Talk:Cardinality
74:Wait a minute!!
63:
56:
55:
33:
32:
26:
2259:
2258:
2254:
2253:
2252:
2250:
2249:
2248:
2224:
2051:
2050:
1904:
1755:
1586:) uses the set
1575:
1565:
1550:
1542:
1421:
1266:wikipedia page!
1147:
1083:choice function
925:
757:, and consider
677:Dedekind finite
535:
506:
497:
476:
451:maximal element
395:
362:
317:right here, in
283:
269:
255:
224:
222:Axiom of Choice
187:
72:
59:
30:
22:
21:
20:
18:Talk:Finite set
12:
11:
5:
2257:
2255:
2210:
2209:
2208:
2207:
2173:
2172:
2171:
2160:
2157:
2154:
2151:
2148:
2145:
2142:
2139:
2136:
2133:
2130:
2127:
2124:
2121:
2118:
2115:
2112:
2109:
2106:
2103:
2100:
2097:
2094:
2091:
2088:
2085:
2082:
2079:
2076:
2073:
2070:
2067:
2064:
2061:
2058:
2034:
2033:
2015:
2012:
1995:
1983:
1982:
1949:
1936:
1935:
1920:
1919:
1903:
1900:
1899:
1898:
1866:
1862:
1861:
1846:
1841:
1840:
1839:
1838:
1826:
1825:
1824:
1806:
1805:
1804:
1770:
1753:
1652:
1640:
1639:
1625:
1622:
1605:
1539:
1538:
1498:
1494:
1493:
1429:
1420:
1417:
1403:
1379:
1371:
1370:
1363:
1362:
1350:Any thoughts?
1348:
1347:
1346:
1332:
1315:
1304:
1301:
1284:
1267:
1244:
1243:
1146:
1143:
1142:
1141:
1067:
1066:
1065:
1064:
1063:
1062:
1061:
1060:
1039:
1038:
1037:
1036:
1035:
1034:
967:
966:
933:
932:
924:
921:
920:
919:
845:, then either
805:in which case
789:, then either
719:
718:
717:
716:
694:
693:
576:in which case
531:
502:
493:
477:be the set of
472:
455:
454:
436:
429:
422:
394:
391:
390:
389:
361:
358:
341:
340:
339:
338:
335:
328:
322:
315:
282:
279:
268:
265:
254:
251:
250:
249:
239:
223:
220:
219:
218:
186:
180:
143:natural number
79:natural number
71:
68:
65:
64:
52:
51:
34:
23:
15:
14:
13:
10:
9:
6:
4:
3:
2:
2256:
2247:
2244:
2240:
2236:
2232:
2228:
2220:
2214:
2206:
2202:
2198:
2194:
2190:
2186:
2182:
2178:
2174:
2149:
2143:
2140:
2137:
2134:
2128:
2125:
2122:
2116:
2113:
2110:
2098:
2095:
2092:
2086:
2083:
2080:
2074:
2071:
2068:
2059:
2049:
2048:
2046:
2042:
2038:
2037:
2036:
2035:
2032:
2028:
2024:
2020:
2016:
2013:
2010:
2009:
2008:
2007:
2003:
1999:
1991:
1987:
1981:
1977:
1973:
1968:
1964:
1963:
1962:
1961:
1957:
1953:
1945:
1942:
1933:
1929:
1925:
1924:
1923:
1917:
1913:
1909:
1908:
1907:
1901:
1897:
1893:
1889:
1885:
1881:
1880:
1879:
1878:
1874:
1870:
1860:
1856:
1852:
1847:
1843:
1842:
1837:
1834:
1831:
1827:
1822:
1818:
1814:
1810:
1809:
1807:
1802:
1798:
1794:
1790:
1789:
1787:
1783:
1779:
1775:
1771:
1768:
1764:
1760:
1756:
1749:
1745:
1741:
1737:
1733:
1729:
1728:
1727:
1723:
1719:
1715:
1711:
1707:
1703:
1699:
1695:
1691:
1687:
1683:
1679:
1675:
1671:
1667:
1666:
1665:
1664:
1660:
1656:
1648:
1644:
1638:
1634:
1630:
1626:
1623:
1620:
1619:
1618:
1617:
1613:
1609:
1601:
1598:
1594:
1589:
1583:
1570:
1562:
1558:
1557:
1549:
1545:
1537:
1533:
1529:
1525:
1521:
1517:
1513:
1512:
1511:
1510:
1506:
1502:
1492:
1488:
1484:
1480:
1476:
1472:
1468:
1464:
1460:
1456:
1452:
1448:
1444:
1443:
1442:
1441:
1437:
1433:
1425:
1418:
1416:
1415:
1411:
1407:
1401:
1397:
1392:
1391:
1387:
1383:
1377:
1368:
1367:
1366:
1361:
1357:
1353:
1349:
1344:
1340:
1336:
1333:
1330:
1326:
1322:
1319:
1318:
1316:
1313:
1309:
1305:
1302:
1299:
1298:
1297:
1296:
1292:
1288:
1280:
1279:
1275:
1271:
1265:
1259:
1258:
1254:
1250:
1242:
1238:
1234:
1231:
1229:
1225:
1220:
1219:
1218:
1217:
1213:
1209:
1205:
1201:
1197:
1193:
1189:
1185:
1181:
1177:
1173:
1169:
1165:
1161:
1158:
1155:
1151:
1144:
1140:
1136:
1132:
1131:TricksterWolf
1128:
1123:
1122:
1121:
1120:
1116:
1112:
1108:
1104:
1100:
1097:from ω into ∪
1096:
1092:
1088:
1084:
1080:
1076:
1072:
1059:
1055:
1051:
1050:TricksterWolf
1047:
1046:
1045:
1044:
1043:
1042:
1041:
1040:
1033:
1029:
1025:
1021:
1017:
1013:
1009:
1005:
1001:
1000:
999:
995:
991:
987:
983:
979:
975:
971:
970:
969:
968:
965:
962:
959:
954:
953:
952:
951:
947:
943:
942:TricksterWolf
938:
930:
929:
928:
922:
918:
914:
910:
906:
905:
904:
903:
899:
895:
890:
888:
884:
880:
876:
872:
868:
864:
860:
856:
852:
848:
844:
840:
836:
832:
828:
824:
820:
816:
813:. Otherwise,
812:
808:
804:
800:
796:
792:
788:
784:
780:
776:
772:
768:
764:
760:
756:
752:
748:
744:
740:
736:
732:
728:
724:
715:
711:
707:
703:
698:
697:
696:
695:
692:
689:
686:
682:
678:
674:
670:
666:
665:
664:
663:
659:
655:
651:
647:
643:
639:
635:
631:
627:
623:
619:
615:
611:
607:
603:
599:
595:
591:
587:
583:
579:
575:
571:
567:
563:
559:
555:
551:
547:
543:
539:
534:
530:
526:
522:
518:
514:
510:
505:
501:
496:
492:
488:
484:
480:
475:
471:
467:
462:
460:
452:
448:
444:
440:
437:
434:
430:
427:
423:
420:
417:
416:
415:
413:
409:
406:
403:
399:
392:
388:
385:
382:
378:
377:
376:
375:
371:
367:
359:
357:
356:
353:
349:
345:
336:
333:
329:
327:
323:
320:
316:
313:
309:
308:
307:
306:
305:
302:
300:
296:
292:
288:
280:
278:
277:
274:
266:
264:
263:
260:
248:
245:
240:
236:
235:
234:
233:
230:
221:
217:
214:
210:
209:
208:
207:
204:
200:
196:
192:
184:
181:
179:
178:
175:
171:
170:{1,2,3,...,n}
167:
162:
160:
156:
152:
148:
144:
139:
137:
132:
130:
126:
122:
119:
116:
113:
110:
108:
103:
101:
96:
94:
89:
86:
84:
80:
75:
69:
62:
58:
57:
49:
45:
41:
40:
35:
28:
27:
19:
2225:— Preceding
2218:
2215:
2211:
2180:
2176:
2044:
2018:
1992:
1988:
1984:
1966:
1946:
1937:
1931:
1927:
1921:
1915:
1911:
1905:
1863:
1830:Arthur Rubin
1820:
1816:
1812:
1800:
1796:
1792:
1785:
1781:
1777:
1766:
1762:
1758:
1751:
1747:
1743:
1739:
1735:
1731:
1713:
1709:
1705:
1701:
1697:
1693:
1689:
1681:
1677:
1673:
1669:
1649:
1645:
1641:
1602:
1596:
1587:
1569:cite journal
1560:
1554:
1544:Lévy, Azriel
1540:
1523:
1519:
1515:
1495:
1478:
1474:
1470:
1466:
1462:
1458:
1454:
1450:
1446:
1426:
1422:
1393:
1372:
1364:
1342:
1338:
1334:
1328:
1324:
1320:
1311:
1307:
1281:
1260:
1245:
1227:
1223:
1221:
1203:
1199:
1195:
1191:
1187:
1183:
1179:
1175:
1171:
1167:
1156:
1148:
1126:
1106:
1102:
1098:
1094:
1090:
1086:
1081:which has a
1078:
1074:
1070:
1068:
1019:
1015:
1011:
1007:
1003:
985:
981:
973:
958:Arthur Rubin
936:
934:
926:
891:
886:
882:
878:
874:
870:
866:
862:
858:
854:
850:
846:
842:
838:
834:
830:
826:
822:
818:
814:
810:
806:
802:
798:
794:
790:
786:
782:
778:
774:
770:
766:
762:
758:
754:
750:
746:
742:
738:
734:
730:
726:
722:
720:
685:Arthur Rubin
668:
649:
645:
641:
637:
633:
629:
625:
621:
617:
613:
609:
605:
601:
597:
593:
589:
585:
581:
577:
573:
569:
565:
561:
557:
553:
549:
545:
541:
537:
532:
528:
524:
520:
516:
512:
508:
503:
499:
494:
490:
486:
482:
473:
469:
465:
463:
458:
456:
446:
438:
432:
425:
418:
404:
398:Arthur Rubin
396:
363:
343:
342:
332:Infinite set
303:
284:
270:
256:
225:
188:
169:
165:
163:
158:
154:
146:
140:
138:May 5, 2005
133:
120:
117:
114:
111:
107:Toby Bartels
104:
99:
97:
90:
87:
76:
73:
60:
43:
37:
1776:implies (∀
449:contains a
319:Cardinality
295:finite sets
229:Fell Collar
91:This isn't
36:This is an
1941:finite set
1400:finite set
1264:finite set
978:Kurt Gödel
616:) and let
366:WardenWalk
360:Redundancy
350:. Regards,
273:Sam Staton
195:finite set
136:bananaclaw
2197:JRSpriggs
2023:JRSpriggs
1972:JRSpriggs
1948:I-finite.
1928:Ia-finite
1912:Ia-finite
1888:JRSpriggs
1851:JRSpriggs
1718:JRSpriggs
1629:JRSpriggs
1578:|ref=harv
1528:JRSpriggs
1483:JRSpriggs
1352:JRSpriggs
1341:)) where
1233:JRSpriggs
1208:JRSpriggs
1111:JRSpriggs
1024:Trovatore
990:JRSpriggs
956:here. —
894:JRSpriggs
877:and thus
706:JRSpriggs
671:; if the
654:JRSpriggs
348:this page
244:Trovatore
213:Trovatore
203:Trovatore
61:Archive 1
2239:contribs
2227:unsigned
2187:and the
1902:Cofinite
1576:Invalid
1546:(1958).
1327:) where
1160:contribs
704:(CC).".
564:, since
408:contribs
379:Done. —
352:Paolo.dL
267:Dedekind
93:Bourbaki
88:-- Anon
1994:matter.
1811:ZF ⊢ (∀
1563:: 1–13.
1182:not in
1178:} (for
679:set is
640:(1)<
636:(0)<
552:. Then
259:Fsmoura
164:As for
141:First,
81:" and "
39:archive
2141:NatNum
2045:NatNum
2043:). If
1833:(talk)
1815:⊆ ZF)(
1795:⊆ ZF)(
1780:⊆ ZF)(
1310:|: -->
961:(talk)
777:, let
769:}. If
688:(talk)
681:finite
544:) for
485:+1 to
199:finite
191:finite
183:Finite
151:finite
83:finite
1551:(PDF)
1337:⊆P(P(
976:. As
909:Cliff
608:+1))=
481:from
100:don't
16:<
2235:talk
2201:talk
2027:talk
2002:talk
1976:talk
1956:talk
1892:talk
1873:talk
1855:talk
1845:VII.
1722:talk
1659:talk
1633:talk
1612:talk
1582:help
1532:talk
1505:talk
1487:talk
1453:and
1436:talk
1410:talk
1386:talk
1356:talk
1291:talk
1274:talk
1253:talk
1237:talk
1212:talk
1154:talk
1135:talk
1129:.)
1115:talk
1054:talk
1028:talk
994:talk
946:talk
913:talk
898:talk
710:talk
658:talk
632:<
548:<
402:talk
381:Emil
370:talk
129:Talk
2219:not
2195:).
1445:If
1323:⊆P(
937:not
885:in
873:in
853:or
833:in
797:or
785:in
749:in
729:of
588:))=
445:of
330:in
324:in
301:!
193:to
166:n=0
147:not
145:is
125:MFH
105:--
2241:)
2237:•
2203:)
2144:
2138:⊨
2132:⇔
2129:ω
2126:∈
2114:∈
2108:∀
2105:⇒
2096:∈
2084:∈
2078:∀
2072:∈
2066:∀
2057:∀
2029:)
2004:)
1996:--
1978:)
1958:)
1950:--
1930:.
1894:)
1875:)
1867:--
1857:)
1828:—
1823:))
1803:))
1791:(∀
1724:)
1661:)
1653:--
1635:)
1614:)
1606:--
1573::
1571:}}
1567:{{
1561:46
1559:.
1553:.
1534:)
1526:.
1507:)
1499:--
1489:)
1438:)
1430:--
1412:)
1404:--
1388:)
1380:--
1358:)
1293:)
1285:--
1276:)
1268:--
1255:)
1239:)
1214:)
1174:∪{
1137:)
1117:)
1085:,
1056:)
1030:)
1004:is
996:)
948:)
915:)
900:)
765:∪{
712:)
660:)
384:J.
372:)
242:--
174:EJ
127::
123:—
2233:(
2199:(
2181:M
2177:V
2159:)
2156:)
2153:)
2150:n
2147:(
2135:M
2123:n
2120:(
2117:M
2111:n
2102:)
2099:M
2093:s
2090:(
2087:r
2081:s
2075:M
2069:r
2063:(
2060:M
2025:(
2019:S
2000:(
1974:(
1967:S
1954:(
1932:S
1926:"
1916:S
1910:"
1890:(
1871:(
1853:(
1821:Q
1817:Q
1813:Q
1801:Q
1797:Q
1793:Q
1786:Q
1782:Q
1778:Q
1767:M
1763:N
1759:N
1754:N
1752:c
1748:c
1744:c
1740:c
1736:N
1732:M
1720:(
1714:V
1710:V
1706:x
1702:x
1698:x
1694:x
1690:x
1682:x
1678:x
1674:x
1670:x
1657:(
1631:(
1610:(
1597:K
1588:K
1584:)
1580:(
1530:(
1524:M
1520:V
1516:M
1503:(
1485:(
1479:V
1475:M
1471:V
1467:M
1463:x
1459:M
1457:∈
1455:x
1451:V
1447:M
1434:(
1408:(
1384:(
1378:.
1354:(
1343:D
1339:D
1335:S
1329:D
1325:D
1321:S
1312:S
1308:S
1289:(
1272:(
1251:(
1235:(
1228:S
1224:S
1210:(
1204:S
1200:S
1196:S
1192:S
1188:S
1184:S
1180:x
1176:x
1172:S
1168:S
1157:·
1152:(
1133:(
1113:(
1107:f
1103:S
1099:S
1095:f
1091:S
1087:B
1079:S
1077:⊆
1075:B
1071:S
1052:(
1026:(
1020:D
1016:D
1012:D
1008:D
992:(
986:D
982:D
974:D
944:(
911:(
896:(
887:T
883:b
879:m
875:S
871:y
867:m
863:S
861:∈
859:y
857:≤
855:b
851:T
847:b
843:b
841:=
839:x
835:S
831:m
827:S
825:∈
823:y
821:≤
819:b
817:≤
815:m
811:T
807:b
803:b
801:≤
799:m
795:T
791:m
787:S
783:x
779:m
775:b
773:≠
771:x
767:b
763:S
761:=
759:T
755:S
751:S
747:x
743:S
739:x
735:m
731:S
727:x
723:S
708:(
656:(
650:g
646:x
642:g
638:g
634:g
630:x
626:l
622:S
618:l
614:n
612:(
610:g
606:n
604:(
602:g
600:(
598:l
594:n
592:(
590:g
586:n
584:(
582:g
580:(
578:h
574:g
570:S
566:h
562:S
558:S
554:g
550:n
546:k
542:k
540:(
538:g
533:n
529:B
527:(
525:f
521:n
519:(
517:g
513:f
509:n
507:|
504:n
500:B
495:n
491:B
487:S
483:n
474:n
470:B
466:S
459:S
453:.
447:S
439:S
433:S
426:S
419:S
405:·
400:(
368:(
321:;
314:;
159:X
155:X
50:.
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.