Theorem dftrrels2 37382

Description: Alternate definition of the class of transitive relations.

I'd prefer to define the class of transitive relations by using the definition of composition by [Suppes] p. 63. df-coSUP (𝐴 ∘ 𝐵) = {⟨𝑥, 𝑦⟩ ∣ ∃𝑢(𝑥𝐴𝑢 ∧ 𝑢𝐵𝑦)} as opposed to the present definition of composition df-co 5683 (𝐴 ∘ 𝐵) = {⟨𝑥, 𝑦⟩ ∣ ∃𝑢(𝑥𝐵𝑢 ∧ 𝑢𝐴𝑦)} because the Suppes definition keeps the order of 𝐴, 𝐵, 𝐶, 𝑅, 𝑆, 𝑇 by default in trsinxpSUP (((𝑅 ∩ (𝐴 × 𝐵)) ∘ (𝑆 ∩ (𝐵 × 𝐶))) ⊆ (𝑇 ∩ (𝐴 × 𝐶)) ↔ ∀𝑥 ∈ 𝐴∀𝑦 ∈ 𝐵∀ 𝑧 ∈ 𝐶((𝑥𝑅𝑦 ∧ 𝑦𝑆𝑧) → 𝑥𝑇𝑧)) while the present definition of composition disarranges them: trsinxp (((𝑆 ∩ (𝐵 × 𝐶)) ∘ (𝑅 ∩ (𝐴 × 𝐵))) ⊆ (𝑇 ∩ (𝐴 × 𝐶 )) ↔ ∀𝑥 ∈ 𝐴∀𝑦 ∈ 𝐵∀𝑧 ∈ 𝐶((𝑥𝑅𝑦 ∧ 𝑦𝑆𝑧) → 𝑥𝑇𝑧) ). This is not mission critical to me, the implication of the Suppes definition is just more aesthetic, at least in the above case.

If we swap to the Suppes definition of class composition, I would define the present class of all transitive sets as df-trsSUP and I would consider to switch the definition of the class of cosets by 𝑅 from the present df-coss 37218 to a df-cossSUP. But perhaps there is a mathematical reason to keep the present definition of composition. (Contributed by Peter Mazsa, 21-Jul-2021.)

Assertion

Ref	Expression
dftrrels2	⊢ TrRels = {𝑟 ∈ Rels ∣ (𝑟 ∘ 𝑟) ⊆ 𝑟}

Proof of Theorem dftrrels2

Step	Hyp	Ref	Expression
1		df-trrels 37380	. 2 ⊢ TrRels = ( Trs ∩ Rels )
2		df-trs 37379	. 2 ⊢ Trs = {𝑟 ∣ ((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) S (𝑟 ∩ (dom 𝑟 × ran 𝑟))}
3		inex1g 5317	. . . . 5 ⊢ (𝑟 ∈ V → (𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∈ V)
4	3	elv 3481	. . . 4 ⊢ (𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∈ V
5		brssr 37308	. . . 4 ⊢ ((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∈ V → (((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) S (𝑟 ∩ (dom 𝑟 × ran 𝑟)) ↔ ((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) ⊆ (𝑟 ∩ (dom 𝑟 × ran 𝑟))))
6	4, 5	ax-mp 5	. . 3 ⊢ (((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) S (𝑟 ∩ (dom 𝑟 × ran 𝑟)) ↔ ((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) ⊆ (𝑟 ∩ (dom 𝑟 × ran 𝑟)))
7		elrels6 37297	. . . . . . 7 ⊢ (𝑟 ∈ V → (𝑟 ∈ Rels ↔ (𝑟 ∩ (dom 𝑟 × ran 𝑟)) = 𝑟))
8	7	elv 3481	. . . . . 6 ⊢ (𝑟 ∈ Rels ↔ (𝑟 ∩ (dom 𝑟 × ran 𝑟)) = 𝑟)
9	8	biimpi 215	. . . . 5 ⊢ (𝑟 ∈ Rels → (𝑟 ∩ (dom 𝑟 × ran 𝑟)) = 𝑟)
10	9, 9	coeq12d 5861	. . . 4 ⊢ (𝑟 ∈ Rels → ((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) = (𝑟 ∘ 𝑟))
11	10, 9	sseq12d 4013	. . 3 ⊢ (𝑟 ∈ Rels → (((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) ⊆ (𝑟 ∩ (dom 𝑟 × ran 𝑟)) ↔ (𝑟 ∘ 𝑟) ⊆ 𝑟))
12	6, 11	bitrid 283	. 2 ⊢ (𝑟 ∈ Rels → (((𝑟 ∩ (dom 𝑟 × ran 𝑟)) ∘ (𝑟 ∩ (dom 𝑟 × ran 𝑟))) S (𝑟 ∩ (dom 𝑟 × ran 𝑟)) ↔ (𝑟 ∘ 𝑟) ⊆ 𝑟))
13	1, 2, 12	abeqinbi 37058	1 ⊢ TrRels = {𝑟 ∈ Rels ∣ (𝑟 ∘ 𝑟) ⊆ 𝑟}

Syntax hints:   ↔ wb 205   = wceq 1542   ∈ wcel 2107  {crab 3433  Vcvv 3475   ∩ cin 3945   ⊆ wss 3946   class class class wbr 5146   × cxp 5672  dom cdm 5674  ran crn 5675   ∘ ccom 5678   Rels crels 36982   S cssr 36983   Trs ctrs 36993   TrRels ctrrels 36994

This theorem was proved from axioms:  ax-mp 5  ax-1 6  ax-2 7  ax-3 8  ax-gen 1798  ax-4 1812  ax-5 1914  ax-6 1972  ax-7 2012  ax-8 2109  ax-9 2117  ax-10 2138  ax-12 2172  ax-ext 2704  ax-sep 5297  ax-nul 5304  ax-pr 5425

This theorem depends on definitions:  df-bi 206  df-an 398  df-or 847  df-3an 1090  df-tru 1545  df-fal 1555  df-ex 1783  df-nf 1787  df-sb 2069  df-clab 2711  df-cleq 2725  df-clel 2811  df-ral 3063  df-rex 3072  df-rab 3434  df-v 3477  df-dif 3949  df-un 3951  df-in 3953  df-ss 3963  df-nul 4321  df-if 4527  df-pw 4602  df-sn 4627  df-pr 4629  df-op 4633  df-br 5147  df-opab 5209  df-xp 5680  df-rel 5681  df-cnv 5682  df-co 5683  df-dm 5684  df-rn 5685  df-res 5686  df-rels 37292  df-ssr 37305  df-trs 37379  df-trrels 37380

This theorem is referenced by:  dftrrels3  37383  eltrrels2  37386  dfeqvrels2  37395