P. 380 of Theorem List - Metamath Proof Explorer

Theorem List for Metamath Proof Explorer - 37901-38000   *Has distinct variable group(s)

Type	Label	Description
Statement
Theorem	bj-pr2val 37901	Value of the second projection. (Contributed by BJ, 6-Apr-2019.)
⊢ pr2 ({𝐴} × tag 𝐵) = if(𝐴 = 1o, 𝐵, ∅)
Theorem	bj-pr22val 37902	Value of the second projection of a couple. (Contributed by BJ, 6-Oct-2018.)
⊢ pr2 ⦅𝐴, 𝐵⦆ = 𝐵
Theorem	bj-pr2ex 37903	Sethood of the second projection. (Contributed by BJ, 6-Oct-2018.)
⊢ (𝐴 ∈ 𝑉 → pr2 𝐴 ∈ V)
Theorem	bj-2uplth 37904	The characteristic property of couples. Note that this holds without sethood hypotheses (compare opth 5445). (Contributed by BJ, 6-Oct-2018.)
⊢ (⦅𝐴, 𝐵⦆ = ⦅𝐶, 𝐷⦆ ↔ (𝐴 = 𝐶 ∧ 𝐵 = 𝐷))
Theorem	bj-2uplex 37905	A couple is a set if and only if its coordinates are sets. For the advantages offered by the reverse closure property, see the section head comment. (Contributed by BJ, 6-Oct-2018.)
⊢ (⦅𝐴, 𝐵⦆ ∈ V ↔ (𝐴 ∈ V ∧ 𝐵 ∈ V))
Theorem	bj-2upln0 37906	A couple is nonempty. (Contributed by BJ, 21-Apr-2019.)
⊢ ⦅𝐴, 𝐵⦆ ≠ ∅
Theorem	bj-2upln1upl 37907	A couple is never equal to a monuple. It is in order to have this "non-clashing" result that tagging was used. Without tagging, we would have ⦅𝐴, ∅⦆ = ⦅𝐴⦆. Note that in the context of Morse tuples, it is natural to define the 0-tuple as the empty set. Therefore, the present theorem together with bj-1upln0 37892 and bj-2upln0 37906 tell us that an m-tuple may equal an n-tuple only when m = n, at least for m, n <= 2, but this result would extend as soon as we define n-tuples for higher values of n. (Contributed by BJ, 21-Apr-2019.)
⊢ ⦅𝐴, 𝐵⦆ ≠ ⦅𝐶⦆
21.19.5.16  Set theory: elementary operations relative to a universe Some elementary set-theoretic operations "relative to a universe" (by which is merely meant some given class considered as a universe).
Theorem	bj-rcleqf 37908	Relative version of cleqf 2951. (Contributed by BJ, 27-Dec-2023.)
⊢ Ⅎ𝑥𝐴    &   ⊢ Ⅎ𝑥𝐵    &   ⊢ Ⅎ𝑥𝑉    ⇒   ⊢ ((𝑉 ∩ 𝐴) = (𝑉 ∩ 𝐵) ↔ ∀𝑥 ∈ 𝑉 (𝑥 ∈ 𝐴 ↔ 𝑥 ∈ 𝐵))
Theorem	bj-rcleq 37909*	Relative version of dfcleq 2754. (Contributed by BJ, 27-Dec-2023.)
⊢ ((𝑉 ∩ 𝐴) = (𝑉 ∩ 𝐵) ↔ ∀𝑥 ∈ 𝑉 (𝑥 ∈ 𝐴 ↔ 𝑥 ∈ 𝐵))
Theorem	bj-reabeq 37910*	Relative form of eqabb 2900. (Contributed by BJ, 27-Dec-2023.)
⊢ ((𝑉 ∩ 𝐴) = {𝑥 ∈ 𝑉 ∣ 𝜑} ↔ ∀𝑥 ∈ 𝑉 (𝑥 ∈ 𝐴 ↔ 𝜑))
Theorem	bj-disj2r 37911	Relative version of ssdifin0 4441, allowing a biconditional, and of disj2 4411. (Contributed by BJ, 11-Nov-2021.) This proof does not rely, even indirectly, on ssdifin0 4441 nor disj2 4411. (Proof modification is discouraged.)
⊢ ((𝐴 ∩ 𝑉) ⊆ (𝑉 ∖ 𝐵) ↔ ((𝐴 ∩ 𝐵) ∩ 𝑉) = ∅)
Theorem	bj-sscon 37912	Contraposition law for relative subclasses. Relative and generalized version of ssconb 4089. Shortens ssconb 4089, conss2 45385. (Contributed by BJ, 11-Nov-2021.) This proof does not rely, even indirectly, on ssconb 4089 nor conss2 45385. (Proof modification is discouraged.)
⊢ ((𝐴 ∩ 𝑉) ⊆ (𝑉 ∖ 𝐵) ↔ (𝐵 ∩ 𝑉) ⊆ (𝑉 ∖ 𝐴))
21.19.5.17  Axioms for finite unions In this section, we introduce the axiom of singleton ax-bj-sn 37916 and the axiom of binary union ax-bj-bun 37920. Both axioms are implied by the standard axioms of unordered pair ax-pr 5391 and of union ax-un 7740 (see snex 5397 and unex 7750). Conversely, the axiom of unordered pair ax-pr 5391 is implied by the axioms of singleton and of binary union, as proved in bj-prexg 37922 and bj-prex 37923. The axioms of union ax-un 7740 and of powerset ax-pow 5327 are independent of these axioms: consider respectively the class of pseudo-hereditarily sets of cardinality less than a given singular strong limit cardinal, see Greg Oman, On the axiom of union, Arch. Math. Logic (2010) 49:283--289 (that model does have finite unions), and the class of well-founded hereditarily countable sets (or hereditarily less than a given uncountable regular cardinal). See also https://mathoverflow.net/questions/81815 5327 and https://mathoverflow.net/questions/48365 5327. A proof by finite induction shows that the existence of finite unions is equivalent to the existence of binary unions and of nullary unions (the latter being the axiom of the empty set ax-nul 5260). The axiom of binary union is useful in theories without the axioms of union ax-un 7740 and of powerset ax-pow 5327. For instance, the class of well-founded sets hereditarily of cardinality at most 𝑛 ∈ ℕ0 with ordinary membership relation is a model of { ax-ext 2733, ax-rep 5232, ax-sep 5249, ax-nul 5260, ax-reg 9570 } and the axioms of existence of unordered 𝑚-tuples for all 𝑚 ≤ 𝑛, and in most cases one would like to rule out such models, hence the need for extra axioms, typically variants of powersets or unions. The axiom of adjunction ax-bj-adj 37925 is more widely used, and is an axiom of General Set Theory. We prove how to retrieve it from binary union and singleton in bj-adjfrombun 37929 and conversely how to prove from adjunction singleton (bj-snfromadj 37927) and unordered pair (bj-prfromadj 37928).
Theorem	bj-abex 37913*	Two ways of stating that the extension of a formula is a set. (Contributed by BJ, 18-Jan-2025.) (Proof modification is discouraged.)
⊢ ({𝑥 ∣ 𝜑} ∈ V ↔ ∃𝑦∀𝑥(𝑥 ∈ 𝑦 ↔ 𝜑))
Theorem	bj-clex 37914*	Two ways of stating that a class is a set. (Contributed by BJ, 18-Jan-2025.) (Proof modification is discouraged.)
⊢ (𝑥 ∈ 𝐴 ↔ 𝜑)    ⇒   ⊢ (𝐴 ∈ V ↔ ∃𝑦∀𝑥(𝑥 ∈ 𝑦 ↔ 𝜑))
Theorem	bj-axsn 37915*	Two ways of stating the axiom of singleton (which is the universal closure of either side, see ax-bj-sn 37916). (Contributed by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ ({𝑥} ∈ V ↔ ∃𝑦∀𝑧(𝑧 ∈ 𝑦 ↔ 𝑧 = 𝑥))
Axiom	ax-bj-sn 37916*	Axiom of singleton. (Contributed by BJ, 12-Jan-2025.)
⊢ ∀𝑥∃𝑦∀𝑧(𝑧 ∈ 𝑦 ↔ 𝑧 = 𝑥)
Theorem	bj-snexg 37917	A singleton built on a set is a set. Contrary to bj-snex 37918, this proof is intuitionistically valid and does not require ax-nul 5260. (Contributed by NM, 7-Aug-1994.) Extract it from snex 5397 and prove it from ax-bj-sn 37916. (Revised by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ (𝐴 ∈ 𝑉 → {𝐴} ∈ V)
Theorem	bj-snex 37918	A singleton is a set. See also snex 5397, snexALT 5345. (Contributed by NM, 7-Aug-1994.) Prove it from ax-bj-sn 37916. (Revised by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ {𝐴} ∈ V
Theorem	bj-axbun 37919*	Two ways of stating the axiom of binary union (which is the universal closure of either side, see ax-bj-bun 37920). (Contributed by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ ((𝑥 ∪ 𝑦) ∈ V ↔ ∃𝑧∀𝑡(𝑡 ∈ 𝑧 ↔ (𝑡 ∈ 𝑥 ∨ 𝑡 ∈ 𝑦)))
Axiom	ax-bj-bun 37920*	Axiom of binary union. (Contributed by BJ, 12-Jan-2025.)
⊢ ∀𝑥∀𝑦∃𝑧∀𝑡(𝑡 ∈ 𝑧 ↔ (𝑡 ∈ 𝑥 ∨ 𝑡 ∈ 𝑦))
Theorem	bj-unexg 37921	Existence of binary unions of sets, proved from ax-bj-bun 37920. (Contributed by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ ((𝐴 ∈ 𝑉 ∧ 𝐵 ∈ 𝑊) → (𝐴 ∪ 𝐵) ∈ V)
Theorem	bj-prexg 37922	Existence of unordered pairs formed on sets, proved from ax-bj-sn 37916 and ax-bj-bun 37920. Contrary to bj-prex 37923, this proof is intuitionistically valid and does not require ax-nul 5260. (Contributed by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ ((𝐴 ∈ 𝑉 ∧ 𝐵 ∈ 𝑊) → {𝐴, 𝐵} ∈ V)
Theorem	bj-prex 37923	Existence of unordered pairs proved from ax-bj-sn 37916 and ax-bj-bun 37920. (Contributed by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ {𝐴, 𝐵} ∈ V
Theorem	bj-axadj 37924*	Two ways of stating the axiom of adjunction (which is the universal closure of either side, see ax-bj-adj 37925). (Contributed by BJ, 12-Jan-2025.) (Proof modification is discouraged.)
⊢ ((𝑥 ∪ {𝑦}) ∈ V ↔ ∃𝑧∀𝑡(𝑡 ∈ 𝑧 ↔ (𝑡 ∈ 𝑥 ∨ 𝑡 = 𝑦)))
Axiom	ax-bj-adj 37925*	Axiom of adjunction. (Contributed by BJ, 19-Jan-2025.)
⊢ ∀𝑥∀𝑦∃𝑧∀𝑡(𝑡 ∈ 𝑧 ↔ (𝑡 ∈ 𝑥 ∨ 𝑡 = 𝑦))
Theorem	bj-adjg1 37926	Existence of the result of the adjunction (generalized only in the first term since this suffices for current applications). (Contributed by BJ, 19-Jan-2025.) (Proof modification is discouraged.)
⊢ (𝐴 ∈ 𝑉 → (𝐴 ∪ {𝑥}) ∈ V)
Theorem	bj-snfromadj 37927	Singleton from adjunction and empty set. (Contributed by BJ, 19-Jan-2025.) (Proof modification is discouraged.)
⊢ {𝑥} ∈ V
Theorem	bj-prfromadj 37928	Unordered pair from adjunction. (Contributed by BJ, 19-Jan-2025.) (Proof modification is discouraged.)
⊢ {𝑥, 𝑦} ∈ V
Theorem	bj-adjfrombun 37929	Adjunction from singleton and binary union. (Contributed by BJ, 19-Jan-2025.) (Proof modification is discouraged.)
⊢ (𝑥 ∪ {𝑦}) ∈ V
21.19.5.18  Set theory: miscellaneous Miscellaneous theorems of set theory.
Theorem	eleq2w2ALT 37930	Alternate proof of eleq2w2 2757 and special instance of eleq2 2850. (Contributed by BJ, 22-Sep-2024.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ (𝐴 = 𝐵 → (𝑥 ∈ 𝐴 ↔ 𝑥 ∈ 𝐵))
Theorem	bj-clel3gALT 37931*	Alternate proof of clel3g 3615. (Contributed by BJ, 1-Sep-2024.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ (𝐵 ∈ 𝑉 → (𝐴 ∈ 𝐵 ↔ ∃𝑥(𝑥 = 𝐵 ∧ 𝐴 ∈ 𝑥)))
Theorem	bj-pw0ALT 37932	Alternate proof of pw0 4773. The proofs have a similar structure: pw0 4773 uses the definitions of powerclass and singleton as class abstractions, whereas bj-pw0ALT 37932 uses characterizations of their elements. Both proofs then use transitivity of a congruence relation (equality for pw0 4773 and biconditional for bj-pw0ALT 37932) to translate the property ss0b 4351 into the wanted result. To translate a biconditional into a class equality, pw0 4773 uses abbii 2828 (which yields an equality of class abstractions), while bj-pw0ALT 37932 uses eqriv 2758 (which requires a biconditional of membership of a given setvar variable). Note that abbii 2828, through its closed form abbi 2826, is proved from eqrdv 2759, which is the deduction form of eqriv 2758. In the other direction, velpw 4562 and velsn 4600 are proved from the definitions of powerclass and singleton using elabg 3630, which is a version of abbii 2828 suited for membership characterizations. (Contributed by BJ, 14-Apr-2024.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ 𝒫 ∅ = {∅}
Theorem	bj-sselpwuni 37933	Quantitative version of ssexg 5281: a subset of an element of a class is an element of the powerclass of the union of that class. (Contributed by BJ, 6-Apr-2024.)
⊢ ((𝐴 ⊆ 𝐵 ∧ 𝐵 ∈ 𝑉) → 𝐴 ∈ 𝒫 ∪ 𝑉)
Theorem	bj-unirel 37934	Quantitative version of uniexr 7766: if the union of a class is an element of a class, then that class is an element of the double powerclass of the union of this class. (Contributed by BJ, 6-Apr-2024.)
⊢ (∪ 𝐴 ∈ 𝑉 → 𝐴 ∈ 𝒫 𝒫 ∪ 𝑉)
Theorem	bj-elpwg 37935	If the intersection of two classes is a set, then inclusion among these classes is equivalent to membership in the powerclass. Common generalization of elpwg 4560 and elpw2g 5295 (the latter of which could be proved from it). (Contributed by BJ, 31-Dec-2023.)
⊢ ((𝐴 ∩ 𝐵) ∈ 𝑉 → (𝐴 ∈ 𝒫 𝐵 ↔ 𝐴 ⊆ 𝐵))
Theorem	bj-velpwALT 37936*	This theorem bj-velpwALT 37936 and the next theorem bj-elpwgALT 37937 are alternate proofs of velpw 4562 and elpwg 4560 respectively, where one proves first the setvar case and then generalizes using vtoclbg 3520 instead of proving first the general case using elab2g 3634 and then specifying. Here, this results in needing an extra DV condition, a longer combined proof and use of ax-12 2213. In other cases, that order is better (e.g., vsnex 5393 proved before snexg 5398). (Contributed by BJ, 17-Jan-2025.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ (𝑥 ∈ 𝒫 𝐴 ↔ 𝑥 ⊆ 𝐴)
Theorem	bj-elpwgALT 37937	Alternate proof of elpwg 4560. See comment for bj-velpwALT 37936. (Contributed by BJ, 17-Jan-2025.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ (𝐴 ∈ 𝑉 → (𝐴 ∈ 𝒫 𝐵 ↔ 𝐴 ⊆ 𝐵))
Theorem	bj-vjust 37938	Justification theorem for dfv2 3454 if it were the definition. See also vjust 3452. (Contributed by BJ, 30-Nov-2019.) (Proof modification is discouraged.)
⊢ {𝑥 ∣ ⊤} = {𝑦 ∣ ⊤}
Theorem	bj-nul 37939*	Two formulations of the axiom of the empty set ax-nul 5260. Proposal: place it right before ax-nul 5260. (Contributed by BJ, 30-Nov-2019.) (Proof modification is discouraged.)
⊢ (∅ ∈ V ↔ ∃𝑥∀𝑦 ¬ 𝑦 ∈ 𝑥)
Theorem	bj-nuliota 37940*	Definition of the empty set using the definite description binder. See also bj-nuliotaALT 37941. (Contributed by BJ, 30-Nov-2019.) (Proof modification is discouraged.)
⊢ ∅ = (℩𝑥∀𝑦 ¬ 𝑦 ∈ 𝑥)
Theorem	bj-nuliotaALT 37941*	Alternate proof of bj-nuliota 37940. Note that this alternate proof uses the fact that ℩𝑥𝜑 evaluates to ∅ when there is no 𝑥 satisfying 𝜑 (iotanul 6511). This is an implementation detail of the encoding currently used in set.mm and should be avoided. (Contributed by BJ, 30-Nov-2019.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ ∅ = (℩𝑥∀𝑦 ¬ 𝑦 ∈ 𝑥)
Theorem	bj-vtoclgfALT 37942	Alternate proof of vtoclgf 3530. Proof from vtoclgft 3516. (This may have been the original proof before shortening.) (Contributed by BJ, 30-Sep-2019.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ Ⅎ𝑥𝐴    &   ⊢ Ⅎ𝑥𝜓    &   ⊢ (𝑥 = 𝐴 → (𝜑 ↔ 𝜓))    &   ⊢ 𝜑    ⇒   ⊢ (𝐴 ∈ 𝑉 → 𝜓)
Theorem	bj-elsn12g 37943	Join of elsng 4598 and elsn2g 4625. (Contributed by BJ, 18-Nov-2023.)
⊢ ((𝐴 ∈ 𝑉 ∨ 𝐵 ∈ 𝑊) → (𝐴 ∈ {𝐵} ↔ 𝐴 = 𝐵))
Theorem	bj-elsnb 37944	Biconditional version of elsng 4598. (Contributed by BJ, 18-Nov-2023.)
⊢ (𝐴 ∈ {𝐵} ↔ (𝐴 ∈ V ∧ 𝐴 = 𝐵))
Theorem	bj-pwcfsdom 37945	Remove hypothesis from pwcfsdom 10649. Illustration of how to remove a "proof-facilitating hypothesis". Shortens theorems using pwcfsdom 10649. (Contributed by BJ, 14-Sep-2019.)
⊢ (ℵ‘𝐴) ≺ ((ℵ‘𝐴) ↑m (cf‘(ℵ‘𝐴)))
Theorem	bj-grur1 37946	Remove hypothesis from grur1 10886. Illustration of how to remove a "definitional hypothesis". This makes its uses longer, but the theorem feels more self-contained. It looks preferable when the defined term appears only once in the conclusion. (Contributed by BJ, 14-Sep-2019.)
⊢ ((𝑈 ∈ Univ ∧ 𝑈 ∈ ∪ (𝑅1 “ On)) → 𝑈 = (𝑅1‘(𝑈 ∩ On)))
Theorem	bj-bm1.3ii 37947*	The extension of a predicate (𝜑(𝑧)) is included in a set (𝑥) if and only if it is a set (𝑦). Sufficiency is obvious, and necessity is the content of the axiom of separation ax-sep 5249. Similar to Theorem 1.3(ii) of [BellMachover] p. 463. (Contributed by NM, 21-Jun-1993.) Generalized to a closed form biconditional with existential quantifications using two different setvars 𝑥, 𝑦 (which need not be disjoint). (Revised by BJ, 8-Aug-2022.) TODO: move after sepexi 5256. Relabel ("sepbi"?).
⊢ (∃𝑥∀𝑧(𝜑 → 𝑧 ∈ 𝑥) ↔ ∃𝑦∀𝑧(𝑧 ∈ 𝑦 ↔ 𝜑))
Theorem	bj-dfid2ALT 37948	Alternate version of dfid2 5548. (Contributed by BJ, 9-Nov-2024.) (Proof modification is discouraged.) Use df-id 5546 instead to make the semantics of the construction df-opab 5168 clearer. (New usage is discouraged.)
⊢ I = {⟨𝑥, 𝑥⟩ ∣ ⊤}
Theorem	bj-0nelopab 37949	The empty set is never an element in an ordered-pair class abstraction. (Contributed by Alexander van der Vekens, 5-Nov-2017.) (Proof shortened by BJ, 22-Jul-2023.) TODO: move to the main section when one can reorder sections so that we can use relopab 5802 (this is a very limited reordering).
⊢ ¬ ∅ ∈ {⟨𝑥, 𝑦⟩ ∣ 𝜑}
Theorem	bj-brrelex12ALT 37950	Two classes related by a binary relation are both sets. Alternate proof of brrelex12 5703. (Contributed by BJ, 14-Jul-2023.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ ((Rel 𝑅 ∧ 𝐴𝑅𝐵) → (𝐴 ∈ V ∧ 𝐵 ∈ V))
Theorem	bj-epelg 37951	The membership relation and the membership predicate agree when the "containing" class is a set. General version of epel 5554 and closed form of epeli 5553. (Contributed by Scott Fenton, 27-Mar-2011.) (Revised by Mario Carneiro, 28-Apr-2015.) TODO: move it to the main section after reordering to have brrelex1i 5707 available. (Proof shortened by BJ, 14-Jul-2023.) (Proof modification is discouraged.)
⊢ (𝐵 ∈ 𝑉 → (𝐴 E 𝐵 ↔ 𝐴 ∈ 𝐵))
Theorem	bj-epelb 37952	Two classes are related by the membership relation if and only if they are related by the membership relation (i.e., the first is an element of the second) and the second is a set (hence so is the first). TODO: move to Main after reordering to have brrelex2i 5708 available. Check if it is shorter to prove bj-epelg 37951 first or bj-epelb 37952 first. (Contributed by BJ, 14-Jul-2023.)
⊢ (𝐴 E 𝐵 ↔ (𝐴 ∈ 𝐵 ∧ 𝐵 ∈ V))
Theorem	bj-nsnid 37953	A set does not contain the singleton formed on it. More precisely, one can prove that a class contains the singleton formed on it if and only if it is proper and contains the empty set (since it is "the singleton formed on" any proper class, see snprc 4678): ⊢ ¬ ({𝐴} ∈ 𝐴 ↔ (∅ ∈ 𝐴 → 𝐴 ∈ V)). (Contributed by BJ, 4-Feb-2023.)
⊢ (𝐴 ∈ 𝑉 → ¬ {𝐴} ∈ 𝐴)
Theorem	bj-rdg0gALT 37954	Alternate proof of rdg0g 8419. More direct since it bypasses tz7.44-1 8398 and rdg0 8413 (and vtoclg 3518, vtoclga 3537). (Contributed by NM, 25-Apr-1995.) More direct proof. (Revised by BJ, 17-Nov-2024.) (Proof modification is discouraged.) (New usage is discouraged.)
⊢ (𝐴 ∈ 𝑉 → (rec(𝐹, 𝐴)‘∅) = 𝐴)
Theorem	bj-vn0ALT 37955	Alternate proof of vn0 4291 which does not use eqabbw 2834 (and is shorter than vn0 4291 when eqabbw 2834 is inlined). (Contributed by BJ, 12-Jul-2026.) Using the same dummy variable for 𝑦 and 𝑧 slightly reduces the proof size. (Proof modification is discouraged.) (New usage is discouraged.)
⊢ V ≠ ∅
21.19.5.19  Axioms of separation and replacement This section proves basic relations among some standard axioms of set theory, in particular the axiom of separation (the universal closure of ax-sep 5249) and the version of the axiom of replacement requiring the functional relation in the axiom to be a (total) function, bj-rep 37957. These axioms often appear (as specific instances) in the hypotheses of the theorems in this section.
Theorem	bj-axnul 37956*	Over the base theory ax-1 6-- ax-5 1943, the axiom of separation implies the weak emptyset axiom. By "weak emptyset axiom", we mean the axiom asserting existence of an empty set (which can be called "the" empty set when the axiom of extensionality ax-ext 2733 is posited) provided existence of a set (the True truth constant existentially quantified over a fresh variable, extru 2008). This is the conclusion of bj-axnul 37956. Note that the weak emptyset axiom implies ⊢ (∃𝑥⊤ → ∃𝑦⊤) without DV conditions hence also the same statement as the weak emptyset axiom without DV conditions on 𝑥, but only on 𝑦, 𝑧. By "axiom of separation", we mean the universal closure of ax-sep 5249, simulated here by its instance with ⊥ substituted for 𝜑 (and with the variable used to assert existence in the weak emptyset axiom substituted for the containing set) as the hypothesis of bj-axnul 37956. In particular, the axiom of existence extru 2008 and the axiom of separation together imply the emptyset axiom (and conversely, the emptyset axiom implies the axiom of existence). Note: this theorem does not require a disjointness condition on 𝑦, 𝑧, although both axioms should be stated with all variables disjoint. This proof only uses an instance of the axiom of separation with a bounded formula, so is valid in a constructive setting (see the CZF section in the "Intuitionistic Logic Explorer" iset.mm). (Contributed by BJ, 8-Mar-2026.) (Proof modification is discouraged.)
⊢ ∀𝑥∃𝑦∀𝑧(𝑧 ∈ 𝑦 ↔ (𝑧 ∈ 𝑥 ∧ ⊥))    ⇒   ⊢ (∃𝑥⊤ → ∃𝑦∀𝑧 ∈ 𝑦 ⊥)
Theorem	bj-rep 37957*	Version of the axiom of replacement requiring the functional relation in the axiom to be a (total) function from ax-rep 5232 (in the form of axrep6 5240). (Contributed by BJ, 14-Mar-2026.) The proof proves the statement without the DV condition on 𝑥, 𝜑, but the DV condition is added to this statement to show that this weaker version is sufficient. (Proof modification is discouraged.) (New usage is discouraged.)
⊢ ∀𝑥(∀𝑦 ∈ 𝑥 ∃!𝑧𝜑 → ∃𝑡∀𝑧(𝑧 ∈ 𝑡 ↔ ∃𝑦 ∈ 𝑥 𝜑))
Theorem	bj-axseprep 37958*	Axiom of separation (universal closure of ax-sep 5249) from a weak form of the axiom of replacement requiring that the functional relation in it be a (total) function and the weak emptyset axiom (existence of an empty set provided existence of a set), as written in the theorem's hypotheses. This result shows that the weak emptyset axiom is not only the result of a cheap way to avoid an axiom redundancy (in this case, the existence axiom extru 2008) by adding it as an antecedent, but also permits to prove nontrivial results that hold in nonnecessarily nonempty universes. This proof is by cases so is not intuitionistic. The statement does not require a nonempty universe; most of the proof does not either, and the parts that do (e.g., near sb8ef 2385 and sbequ12r 2288 and eueq2 3668) could be reworked to avoid it. Proof modifications should not introduce steps relying on a nonempty universe, like alrimiv 1960. (Contributed by BJ, 14-Mar-2026.) (Proof modification is discouraged.)
⊢ (∃𝑥⊤ → ∃𝑦∀𝑧 ∈ 𝑦 ⊥)    &   ⊢ ∀𝑥(∀𝑧 ∈ 𝑥 ∃!𝑡𝜓 → ∃𝑦∀𝑡(𝑡 ∈ 𝑦 ↔ ∃𝑧 ∈ 𝑥 𝜓))    &   ⊢ (𝜓 ↔ ((𝜑 ∧ 𝑡 = 𝑧) ∨ (¬ 𝜑 ∧ 𝑡 = 𝑎)))    ⇒   ⊢ ∀𝑥∃𝑦∀𝑧(𝑧 ∈ 𝑦 ↔ (𝑧 ∈ 𝑥 ∧ 𝜑))
Theorem	bj-axreprepsep 37959*	Strong axiom of replacement (universal closure of ax-rep 5232) from the axioms of separation and replacement as written in the theorem's hypotheses. The statement does not require a nonempty universe; most of the proof does not either, except for the use of 19.8a 2218, which could be removed by reworking the proof, since it is applied in a subexpression bound by the variable it introduces. Proof modifications should not introduce steps relying on a nonempty universe, like alrimiv 1960. (Contributed by BJ, 14-Mar-2026.) (Proof modification is discouraged.)
⊢ ∀𝑥∃𝑠∀𝑦(𝑦 ∈ 𝑠 ↔ (𝑦 ∈ 𝑥 ∧ ∃𝑧𝜑))    &   ⊢ ∀𝑠(∀𝑦 ∈ 𝑠 ∃!𝑧𝜑 → ∃𝑡∀𝑧(𝑧 ∈ 𝑡 ↔ ∃𝑦 ∈ 𝑠 𝜑))    ⇒   ⊢ ∀𝑥(∀𝑦 ∈ 𝑥 ∃*𝑧𝜑 → ∃𝑡∀𝑧(𝑧 ∈ 𝑡 ↔ ∃𝑦 ∈ 𝑥 𝜑))
21.19.5.20  Evaluation at a class This section treats the existing predicate Slot (df-slot 17340) as "evaluation at a class" and for the moment does not introduce new syntax for it.
Theorem	bj-evaleq 37960	Equality theorem for the Slot construction. This is currently a duplicate of sloteq 17341 but may diverge from it if/when a token Eval is introduced for evaluation in order to separate it from Slot and any of its possible modifications. (Contributed by BJ, 27-Dec-2021.) (Proof modification is discouraged.)
⊢ (𝐴 = 𝐵 → Slot 𝐴 = Slot 𝐵)
Theorem	bj-evalfun 37961	The evaluation at a class is a function. (Contributed by BJ, 27-Dec-2021.)
⊢ Fun Slot 𝐴
Theorem	bj-evalfn 37962	The evaluation at a class is a function on the universal class. (General form of slotfn 17342). (Contributed by Mario Carneiro, 22-Sep-2015.) (Revised by BJ, 27-Dec-2021.)
⊢ Slot 𝐴 Fn V
Theorem	bj-evalf 37963	The evaluation at a class is a function from the universal class into the universal class. (Contributed by BJ, 17-Mar-2026.)
⊢ Slot 𝐴:V⟶V
Theorem	bj-evalval 37964	Value of the evaluation at a class. Closed form of strfvnd 17343 and strfvn 17344. (Contributed by NM, 9-Sep-2011.) (Revised by Mario Carneiro, 15-Nov-2014.) (Revised by BJ, 27-Dec-2021.)
⊢ (𝐹 ∈ 𝑉 → (Slot 𝐴‘𝐹) = (𝐹‘𝐴))
Theorem	bj-evalid 37965	The evaluation at a set of the identity function is that set. General form of ndxarg 17354. The restriction to a set 𝑉 is necessary since the argument of the function Slot 𝐴 (like that of any function) has to be a set for the evaluation to be meaningful. (Contributed by BJ, 27-Dec-2021.)
⊢ ((𝑉 ∈ 𝑊 ∧ 𝐴 ∈ 𝑉) → (Slot 𝐴‘( I ↾ 𝑉)) = 𝐴)
Theorem	bj-ndxarg 37966	Proof of ndxarg 17354 from bj-evalid 37965. (Contributed by BJ, 27-Dec-2021.) (Proof modification is discouraged.)
⊢ 𝐸 = Slot 𝑁    &   ⊢ 𝑁 ∈ ℕ    ⇒   ⊢ (𝐸‘ndx) = 𝑁
Theorem	bj-evalidval 37967	Closed general form of strndxid 17356. Both sides are equal to (𝐹‘𝐴) by bj-evalid 37965 and bj-evalval 37964 respectively, but bj-evalidval 37967 adds something to bj-evalid 37965 and bj-evalval 37964 in that Slot 𝐴 appears on both sides. (Contributed by BJ, 27-Dec-2021.)
⊢ ((𝑉 ∈ 𝑊 ∧ 𝐴 ∈ 𝑉 ∧ 𝐹 ∈ 𝑈) → (𝐹‘(Slot 𝐴‘( I ↾ 𝑉))) = (Slot 𝐴‘𝐹))
21.19.5.21  Elementwise operations
Syntax	celwise 37968	Syntax for elementwise operations.
class elwise
Definition	df-elwise 37969*	Define the elementwise operation associated with a given operation. For instance, + is the addition of complex numbers (axaddf 11211), so if 𝐴 and 𝐵 are sets of complex numbers, then (𝐴(elwise‘ + )𝐵) is the set of numbers of the form (𝑥 + 𝑦) with 𝑥 ∈ 𝐴 and 𝑦 ∈ 𝐵. The set of odd natural numbers is (({2}(elwise‘ · )ℕ0)(elwise‘ + ){1}), or less formally 2ℕ0 + 1. (Contributed by BJ, 22-Dec-2021.)
⊢ elwise = (𝑜 ∈ V ↦ (𝑥 ∈ V, 𝑦 ∈ V ↦ {𝑧 ∣ ∃𝑢 ∈ 𝑥 ∃𝑣 ∈ 𝑦 𝑧 = (𝑢𝑜𝑣)}))
21.19.5.22  Elementwise intersection (families of sets induced on a subset) Many kinds of structures are given by families of subsets of a given set: Moore collections (df-mre 17736), topologies (df-top 23192), pi-systems, rings of sets, delta-rings, lambda-systems/Dynkin systems, algebras/fields of sets, sigma-algebras/sigma-fields/tribes (df-siga 34723), sigma rings, monotone classes, matroids/independent sets, bornologies, filters. There is a natural notion of structure induced on a subset. It is often given by an elementwise intersection, namely, the family of intersections of sets in the original family with the given subset. In this subsection, we define this notion and prove its main properties. Classical conditions on families of subsets include being nonempty, containing the whole set, containing the empty set, being stable under unions, intersections, subsets, supersets, (relative) complements. Therefore, we prove related properties for the elementwise intersection. We will call (𝑋 ↾t 𝐴) the elementwise intersection on the family 𝑋 by the class 𝐴. REMARK: many theorems are already in set.mm: "MM> SEARCH *rest* / JOIN".
Theorem	bj-rest00 37970	An elementwise intersection on the empty family is the empty set. TODO: this is 0rest 17580. (Contributed by BJ, 27-Apr-2021.)
⊢ (∅ ↾t 𝐴) = ∅
Theorem	bj-restsn 37971	An elementwise intersection on the singleton on a set is the singleton on the intersection by that set. Generalization of bj-restsn0 37974 and bj-restsnid 37976. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑌 ∈ 𝑉 ∧ 𝐴 ∈ 𝑊) → ({𝑌} ↾t 𝐴) = {(𝑌 ∩ 𝐴)})
Theorem	bj-restsnss 37972	Special case of bj-restsn 37971. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑌 ∈ 𝑉 ∧ 𝐴 ⊆ 𝑌) → ({𝑌} ↾t 𝐴) = {𝐴})
Theorem	bj-restsnss2 37973	Special case of bj-restsn 37971. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝐴 ∈ 𝑉 ∧ 𝑌 ⊆ 𝐴) → ({𝑌} ↾t 𝐴) = {𝑌})
Theorem	bj-restsn0 37974	An elementwise intersection on the singleton on the empty set is the singleton on the empty set. Special case of bj-restsn 37971 and bj-restsnss2 37973. TODO: this is restsn 23468. (Contributed by BJ, 27-Apr-2021.)
⊢ (𝐴 ∈ 𝑉 → ({∅} ↾t 𝐴) = {∅})
Theorem	bj-restsn10 37975	Special case of bj-restsn 37971, bj-restsnss 37972, and bj-rest10 37977. (Contributed by BJ, 27-Apr-2021.)
⊢ (𝑋 ∈ 𝑉 → ({𝑋} ↾t ∅) = {∅})
Theorem	bj-restsnid 37976	The elementwise intersection on the singleton on a class by that class is the singleton on that class. Special case of bj-restsn 37971 and bj-restsnss 37972. (Contributed by BJ, 27-Apr-2021.)
⊢ ({𝐴} ↾t 𝐴) = {𝐴}
Theorem	bj-rest10 37977	An elementwise intersection on a nonempty family by the empty set is the singleton on the empty set. TODO: this generalizes rest0 23467 and could replace it. (Contributed by BJ, 27-Apr-2021.)
⊢ (𝑋 ∈ 𝑉 → (𝑋 ≠ ∅ → (𝑋 ↾t ∅) = {∅}))
Theorem	bj-rest10b 37978	Alternate version of bj-rest10 37977. (Contributed by BJ, 27-Apr-2021.)
⊢ (𝑋 ∈ (𝑉 ∖ {∅}) → (𝑋 ↾t ∅) = {∅})
Theorem	bj-restn0 37979	An elementwise intersection on a nonempty family is nonempty. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑋 ∈ 𝑉 ∧ 𝐴 ∈ 𝑊) → (𝑋 ≠ ∅ → (𝑋 ↾t 𝐴) ≠ ∅))
Theorem	bj-restn0b 37980	Alternate version of bj-restn0 37979. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑋 ∈ (𝑉 ∖ {∅}) ∧ 𝐴 ∈ 𝑊) → (𝑋 ↾t 𝐴) ≠ ∅)
Theorem	bj-restpw 37981	The elementwise intersection on a powerset is the powerset of the intersection. This allows to prove for instance that the topology induced on a subset by the discrete topology is the discrete topology on that subset. See also restdis 23476 (which uses distop 23293 and restopn2 23475). (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑌 ∈ 𝑉 ∧ 𝐴 ∈ 𝑊) → (𝒫 𝑌 ↾t 𝐴) = 𝒫 (𝑌 ∩ 𝐴))
Theorem	bj-rest0 37982	An elementwise intersection on a family containing the empty set contains the empty set. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑋 ∈ 𝑉 ∧ 𝐴 ∈ 𝑊) → (∅ ∈ 𝑋 → ∅ ∈ (𝑋 ↾t 𝐴)))
Theorem	bj-restb 37983	An elementwise intersection by a set on a family containing a superset of that set contains that set. (Contributed by BJ, 27-Apr-2021.)
⊢ (𝑋 ∈ 𝑉 → ((𝐴 ⊆ 𝐵 ∧ 𝐵 ∈ 𝑋) → 𝐴 ∈ (𝑋 ↾t 𝐴)))
Theorem	bj-restv 37984	An elementwise intersection by a subset on a family containing the whole set contains the whole subset. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝐴 ⊆ ∪ 𝑋 ∧ ∪ 𝑋 ∈ 𝑋) → 𝐴 ∈ (𝑋 ↾t 𝐴))
Theorem	bj-resta 37985	An elementwise intersection by a set on a family containing that set contains that set. (Contributed by BJ, 27-Apr-2021.)
⊢ (𝑋 ∈ 𝑉 → (𝐴 ∈ 𝑋 → 𝐴 ∈ (𝑋 ↾t 𝐴)))
Theorem	bj-restuni 37986	The union of an elementwise intersection by a set is equal to the intersection with that set of the union of the family. See also restuni 23460 and restuni2 23465. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑋 ∈ 𝑉 ∧ 𝐴 ∈ 𝑊) → ∪ (𝑋 ↾t 𝐴) = (∪ 𝑋 ∩ 𝐴))
Theorem	bj-restuni2 37987	The union of an elementwise intersection on a family of sets by a subset is equal to that subset. See also restuni 23460 and restuni2 23465. (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝑋 ∈ 𝑉 ∧ 𝐴 ⊆ ∪ 𝑋) → ∪ (𝑋 ↾t 𝐴) = 𝐴)
Theorem	bj-restreg 37988	A reformulation of the axiom of regularity using elementwise intersection. (RK: might have to be placed later since theorems in this section are to be moved early (in the section related to the algebra of sets).) (Contributed by BJ, 27-Apr-2021.)
⊢ ((𝐴 ∈ 𝑉 ∧ 𝐴 ≠ ∅) → ∅ ∈ (𝐴 ↾t 𝐴))
21.19.5.23  Moore collections (complements)
Theorem	bj-raldifsn 37989*	All elements in a set satisfy a given property if and only if all but one satisfy that property and that one also does. Typically, this can be used for characterizations that are proved using different methods for a given element and for all others, for instance zero and nonzero numbers, or the empty set and nonempty sets. (Contributed by BJ, 7-Dec-2021.)
⊢ (𝑥 = 𝐵 → (𝜑 ↔ 𝜓))    ⇒   ⊢ (𝐵 ∈ 𝐴 → (∀𝑥 ∈ 𝐴 𝜑 ↔ (∀𝑥 ∈ (𝐴 ∖ {𝐵})𝜑 ∧ 𝜓)))
Theorem	bj-0int 37990*	If 𝐴 is a collection of subsets of 𝑋, like a Moore collection or a topology, two equivalent ways to say that arbitrary intersections of elements of 𝐴 relative to 𝑋 belong to some class 𝐵: the LHS singles out the empty intersection (the empty intersection relative to 𝑋 is 𝑋 and the intersection of a nonempty family of subsets of 𝑋 is included in 𝑋, so there is no need to intersect it with 𝑋). In typical applications, 𝐵 is 𝐴 itself. (Contributed by BJ, 7-Dec-2021.)
⊢ (𝐴 ⊆ 𝒫 𝑋 → ((𝑋 ∈ 𝐵 ∧ ∀𝑥 ∈ (𝒫 𝐴 ∖ {∅})∩ 𝑥 ∈ 𝐵) ↔ ∀𝑥 ∈ 𝒫 𝐴(𝑋 ∩ ∩ 𝑥) ∈ 𝐵))
Theorem	bj-mooreset 37991*	A Moore collection is a set. Therefore, the class Moore of all Moore sets defined in df-bj-moore 37993 is actually the class of all Moore collections. This is also illustrated by the lack of sethood condition in bj-ismoore 37994. Note that the closed sets of a topology form a Moore collection, so a topology is a set, and this remark also applies to many other families of sets (namely, as soon as the whole set is required to be a set of the family, then the associated kind of family has no proper classes: that this condition suffices to impose sethood can be seen in this proof, which relies crucially on uniexr 7766). Note: if, in the above predicate, we substitute 𝒫 𝑋 for 𝐴, then the last ∈ 𝒫 𝑋 could be weakened to ⊆ 𝑋, and then the predicate would be obviously satisfied since ⊢ ∪ 𝒫 𝑋 = 𝑋 (unipw 5418), making 𝒫 𝑋 a Moore collection in this weaker sense, for any class 𝑋, even proper, but the addition of this single case does not add anything interesting. Instead, we have the biconditional bj-discrmoore 38000. (Contributed by BJ, 8-Dec-2021.)
⊢ (∀𝑥 ∈ 𝒫 𝐴(∪ 𝐴 ∩ ∩ 𝑥) ∈ 𝐴 → 𝐴 ∈ V)
Syntax	cmoore 37992	Syntax for the class of Moore collections.
class Moore
Definition	df-bj-moore 37993*	Define the class of Moore collections. This is indeed the class of all Moore collections since these all are sets, as proved in bj-mooreset 37991, and as illustrated by the lack of sethood condition in bj-ismoore 37994. This is to df-mre 17736 (defining Moore) what df-top 23192 (defining Top) is to df-topon 23209 (defining TopOn). For the sake of consistency, the function defined at df-mre 17736 should be denoted by "MooreOn". Note: df-mre 17736 singles out the empty intersection. This is not necessary. It could be written instead ⊢ Moore = (𝑥 ∈ V ↦ {𝑦 ∈ 𝒫 𝒫 𝑥 ∣ ∀𝑧 ∈ 𝒫 𝑦(𝑥 ∩ ∩ 𝑧) ∈ 𝑦}) and the equivalence of both definitions is proved by bj-0int 37990. There is no added generality in defining a "Moore predicate" for arbitrary classes, since a Moore class satisfying such a predicate is automatically a set (see bj-mooreset 37991). TODO: move to the main section. For many families of sets, one can define both the function associating to each set the set of families of that kind on it (like df-mre 17736 and df-topon 23209) or the class of all families of that kind, independent of a base set (like df-bj-moore 37993 or df-top 23192). In general, the former will be more useful and the extra generality of the latter is not necessary. Moore collections, however, are particular in that they are more ubiquitous and are used in a wide variety of applications (for many families of sets, the family of families of a given kind is often a Moore collection, for instance). Therefore, in the case of Moore families, having both definitions is useful. (Contributed by BJ, 27-Apr-2021.)
⊢ Moore = {𝑥 ∣ ∀𝑦 ∈ 𝒫 𝑥(∪ 𝑥 ∩ ∩ 𝑦) ∈ 𝑥}
Theorem	bj-ismoore 37994*	Characterization of Moore collections. Note that there is no sethood hypothesis on 𝐴: it is implied by either side (this is obvious for the LHS, and is the content of bj-mooreset 37991 for the RHS). (Contributed by BJ, 9-Dec-2021.)
⊢ (𝐴 ∈ Moore ↔ ∀𝑥 ∈ 𝒫 𝐴(∪ 𝐴 ∩ ∩ 𝑥) ∈ 𝐴)
Theorem	bj-ismoored0 37995	Necessary condition to be a Moore collection. (Contributed by BJ, 9-Dec-2021.)
⊢ (𝐴 ∈ Moore → ∪ 𝐴 ∈ 𝐴)
Theorem	bj-ismoored 37996	Necessary condition to be a Moore collection. (Contributed by BJ, 9-Dec-2021.)
⊢ (𝜑 → 𝐴 ∈ Moore)    &   ⊢ (𝜑 → 𝐵 ⊆ 𝐴)    ⇒   ⊢ (𝜑 → (∪ 𝐴 ∩ ∩ 𝐵) ∈ 𝐴)
Theorem	bj-ismoored2 37997	Necessary condition to be a Moore collection. (Contributed by BJ, 9-Dec-2021.)
⊢ (𝜑 → 𝐴 ∈ Moore)    &   ⊢ (𝜑 → 𝐵 ⊆ 𝐴)    &   ⊢ (𝜑 → 𝐵 ≠ ∅)    ⇒   ⊢ (𝜑 → ∩ 𝐵 ∈ 𝐴)
Theorem	bj-ismooredr 37998*	Sufficient condition to be a Moore collection. Note that there is no sethood hypothesis on 𝐴: it is a consequence of the only hypothesis. (Contributed by BJ, 9-Dec-2021.)
⊢ ((𝜑 ∧ 𝑥 ⊆ 𝐴) → (∪ 𝐴 ∩ ∩ 𝑥) ∈ 𝐴)    ⇒   ⊢ (𝜑 → 𝐴 ∈ Moore)
Theorem	bj-ismooredr2 37999*	Sufficient condition to be a Moore collection (variant of bj-ismooredr 37998 singling out the empty intersection). Note that there is no sethood hypothesis on 𝐴: it is a consequence of the first hypothesis. (Contributed by BJ, 9-Dec-2021.)
⊢ (𝜑 → ∪ 𝐴 ∈ 𝐴)    &   ⊢ ((𝜑 ∧ (𝑥 ⊆ 𝐴 ∧ 𝑥 ≠ ∅)) → ∩ 𝑥 ∈ 𝐴)    ⇒   ⊢ (𝜑 → 𝐴 ∈ Moore)
Theorem	bj-discrmoore 38000	The powerclass 𝒫 𝐴 is a Moore collection if and only if 𝐴 is a set. It is then called the discrete Moore collection. (Contributed by BJ, 9-Dec-2021.)
⊢ (𝐴 ∈ V ↔ 𝒫 𝐴 ∈ Moore)