OctoInformBeta

OpenAI preprint, September 2026 Β· Β§8. Relational coding of the diagonal argument

Lemma 49: Goodness, predicate tests and the three tagged copies in the relational diagonal construction

The specialization recovers structural observations.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

The relational diagonal construction uses a data parameter to select conditions on predicates. Here that parameter is fixed by seven structural properties of predicates on a carrier with three tagged copies. The formulas concern labelled terms in a pure type system P=(S,π’œ,β„›)P=(S,\mathcal A,\mathcal R), with arbitrary axiom and product-rule relations, a context Ξ“\Gamma, and an assignment Ξ³\gamma of data and proof modes.

Fix a logical calculus LL with falsity βŠ₯\bot, incoming and outgoing proof paths, and universal quantifiers KYK_Y and KPK_P from its quantifier routes. Its implication applies the outgoing path to the proof-to-proof arrow from the incoming wrapper of the antecedent to the consequent. Negation is implication to βŠ₯\bot; conjunction is Β¬(Aβ‡’Β¬B)\neg(A\Rightarrow\neg B)✦, disjunction is Β¬(Β¬A∧¬B)\neg(\neg A\wedge\neg B)✦, and A↔BA\leftrightarrow B is the conjunction of the two implications.✦ Existential quantification is Β¬βˆ€Β¬\neg\forall\neg.✦ Element quantification uses KYK_Y and predicate quantification uses KPK_P; universal quantification applies its incoming proof path beneath a data-to-proof dependent product and its outgoing path outside it.✦ The logical conditional with guard GG and branches A,BA,B is (G∧A)∨(Β¬G∧B)(G\wedge A)\vee(\neg G\wedge B).

A formula AA is inhabited in proof mode in context Ξ“\Gamma under the mode assignment Ξ³\gamma, written Ξ“βŠ©Ξ³A\Gamma\Vdash_\gamma A, when it has an inhabitant satisfying the proof mode discipline.✦ Formula equivalence A≃Γ,Ξ³BA\simeq_{\Gamma,\gamma}B between formulas AA and BB in context Ξ“\Gamma under the mode assignment Ξ³\gamma means uniform typed open transformers in both directions, with a temporary proof input: Ξ“βŠ©Ξ³A⇔some proof-mode p satisfies Ξ“βŠ’p:A,A≃Γ,Ξ³B⇔{some proof-mode t satisfies Ξ“,a:A⊒t:B,some proof-mode u satisfies Ξ“,b:B⊒u:A.\begin{aligned} \Gamma\Vdash_\gamma A &\quad\Longleftrightarrow\quad \text{some proof-mode }p\text{ satisfies }\Gamma\vdash p:A,\\ A\simeq_{\Gamma,\gamma}B &\quad\Longleftrightarrow\quad \begin{cases} \text{some proof-mode }t\text{ satisfies }\Gamma,a:A\vdash t:B,\\ \text{some proof-mode }u\text{ satisfies }\Gamma,b:B\vdash u:A. \end{cases} \end{aligned}(1) The inputs aa and bb in (1) have proof mode. The mode discipline for labelled terms under Ξ³\gamma gives each variable its assigned mode and each sort data mode, and an application has its labelled result mode when its function has that mode and its argument has the labelled argument mode. An abstraction has its labelled result mode when its annotation has data mode and its body has the result mode under extension by the argument mode; a dependent product has data mode when its domain and codomain have data mode, checking the codomain under that same extension.

Channels and Tagged Observations

A specialized relational coding ℬ\mathcal B fixes an observation system and a type A0A_0.✦ Its underlying relational coding has a common probe type UU, an initial data path Ξ±\alpha, and odd retained data paths β„“\ell, rr and dd for logical, raw and predicate channels. A retained data path is generated by identity, positive extensions with a fixed domain and dummy value, and negative extensions with a data tail; its parity counts negative extensions.✦ Positive extensions form data arrows into the previous wrapped type, and negative extensions form data arrows from that type to the tail carrier. A double channel consists of two odd paths.✦

Let OinO_{\mathrm{in}} and OoutO_{\mathrm{out}} be the inner and outer double channels of the observation system. The specialized base type TT, logical probe element type XX, tagged observation type YY, and predicate types are T=Ξ±(A0),X=β„“(U),Y=Oout(Oin(r(U))),𝒫(V)=d(r(V)).T=\alpha(A_0),\qquad X=\ell(U),\qquad Y=O_{\mathrm{out}}\bigl(O_{\mathrm{in}}(r(U))\bigr),\qquad \mathcal P(V)=d(r(V)).(2) Fix a seed sort for the probe. The common probe type UU is a data dependent product over a positive wrapper of that sort. At a type parameter, its codomain is a wrapped function from a single-bundle type to a data tail carrier.✦ The tail encodes formulas by its encoding proof path followed by positive injection✦, and reads them by positive projection followed by its decoding proof path.✦

Reencoding between odd paths builds a term along the target path from the observation associated with the source path.✦ Write raw⁡V\operatorname{raw}_V for reencoding from β„“\ell to rr, and log⁡V\operatorname{log}_V for the reverse reencoding. For a type VV, channel evaluation takes a predicate p:𝒫(V)p:\mathcal P(V) and an element y:β„“(V)y:\ell(V), and the channel roundtrip takes an element y:β„“(V)y:\ell(V); their definitions are ev⁡V(p,y)=observe⁡d(r(V),p,raw⁡V(y)),rt⁡V(y)=log⁡V(raw⁡V(y)).\operatorname{ev}_V(p,y) =\operatorname{observe}_d\bigl(r(V),p,\operatorname{raw}_V(y)\bigr), \qquad \operatorname{rt}_V(y)=\operatorname{log}_V(\operatorname{raw}_V(y)).(3) The observation in (3) instantiates the supplied body at an identity path, applies the retained dummy at a positive step, and at a negative step builds the argument along the preceding path and decodes the resulting application through the tail.✦ For a type VV and a formula-valued body QQ on β„“(V)\ell(V), the encoded predicate Pred⁡V(Q)\operatorname{Pred}_V(Q) is a predicate in 𝒫(V)\mathcal P(V) built along dd over r(V)r(V) from QQ composed with reencoding from rr to β„“\ell.✦ Write e(v,y)=ev⁡T(v,y)e(v,y)=\operatorname{ev}_T(v,y) for specialized evaluation.✦ The data path builder along a retained path returns its supplied term at the identity path and forms a constant data abstraction at a positive extension. At a negative extension it abstracts over the preceding wrapped type, observes that input along the preceding path against the supplied term, and encodes the resulting formula through the tail.

For t∈{0,1,2}t\in\{0,1,2\} and x:Xx:X, the tagged copy ΞΉt(x)\iota_t(x) is an element of YY obtained by first reencoding xx from β„“\ell to rr at UU. It then uses signed payload embedding along the inner double channel with sign true exactly at t=1t=1, and along the outer double channel with sign true exactly at t=2t=2. Signed embedding builds along the second path from the first path’s contravariant observation, keeping that formula for a true sign and negating it for a false sign.✦ Signed recovery uses the observed marker bit to choose between evaluation and its negation.✦ Double channel evaluation at a type VV observes its input along the second path against the term built along the first path from the supplied predicate body. The iterated signed embedding of a term aa at a type VV along a finite list of signs leaves aa unchanged for the empty list. For a list beginning with bb, it first embeds along the remaining signs and then applies signed embedding with sign bb at the carrier obtained by one double channel iteration for each remaining sign. The iterated recovery predicate of a predicate body QQ at depth zero is QQ. At depth n+1n+1, it performs signed recovery at the carrier obtained by nn double channel iterations, using the depth nn recovery predicate as its test.

The query bit string of an index q∈{0,…,8}q\in\{0,\ldots,8\} is the list of its binary digits at positions zero through three, in that order.✦ The nonzero query index of an observation tag t∈{0,…,7}t\in\{0,\ldots,7\} is t+1∈{0,…,8}t+1\in\{0,\ldots,8\}.✦ The query tagged embedding of a term aa at a type VV with query index qq is its iterated signed embedding along the query bit string of qq.✦ The tag predicate for an empty list of signs is logical truth. For a list beginning with bb, it conjoins the marker bit when bb is true, or its negation when bb is false, with signed recovery of the tag predicate for the remaining signs, both at the carrier obtained by one double channel iteration for each remaining sign.

Let bin(y)b_{\mathrm{in}}(y) be the inner marker recovered through the outer channel, and let bout(y)b_{\mathrm{out}}(y) be the outer marker. For a predicate i:𝒫(U)i:\mathcal P(U), let ziz_i denote its uniform predicate probe. A specialized probe observation o(i,y)o(i,y) of y:Yy:Y recovers through both channels and tests the recovered payload with ziz_i.✦ Tagged observational equivalence AE⁡(y,z)\operatorname{AE}(y,z) of elements y,z:Yy,z:Y is the formula AE⁡(y,z)=(bin(y)↔bin(z))∧((bout(y)↔bout(z))βˆ§βˆ€i:𝒫(U),(o(i,y)↔o(i,z))).\operatorname{AE}(y,z)= \bigl(b_{\mathrm{in}}(y)\leftrightarrow b_{\mathrm{in}}(z)\bigr) \wedge\Bigl( \bigl(b_{\mathrm{out}}(y)\leftrightarrow b_{\mathrm{out}}(z)\bigr) \wedge\forall i:\mathcal P(U),\, \bigl(o(i,y)\leftrightarrow o(i,z)\bigr)\Bigr).(4)✦

Probe Order and Structural Properties

The uniform predicate probe ziz_i of a predicate i:𝒫(U)i:\mathcal P(U) is a data abstraction over the lifted type parameter, followed by a wrapped abstraction over its single-bundle argument. Relational element lifting sends an element of the relational base type to XX by building along the logical path at UU from its section pair query as a function of the probe.✦ The body of ziz_i uses recovery access to the predicate pulled back along this lifting, at the positive projection of the type parameter. Recovery access retains the dummy parameter test as its guard. For a predicate i:𝒫(U)i:\mathcal P(U) and an element x:Xx:X, the probe order iβ‰Όxi\preceq x is the formula obtained by observing xx along the logical path at UU against ziz_i.✦ Predicate equivalence iβ‰ˆπ’«iβ€²i\approx_{\mathcal P}i' for i,iβ€²:𝒫(U)i,i':\mathcal P(U), probe element equivalence xβ‰ˆXzx\approx_X z for x,z:Xx,z:X, extensionality Ext⁡(i)\operatorname{Ext}(i) of a predicate i:𝒫(U)i:\mathcal P(U), and order validity Valid⁡(x)\operatorname{Valid}(x) of an element x:Xx:X are the formulas iβ‰ˆπ’«iβ€²=βˆ€x:X,(ev⁡U(i,x)↔ev⁡U(iβ€²,x)),xβ‰ˆXz=βˆ€i:𝒫(U),(iβ‰Όx↔iβ‰Όz),Ext⁡(i)=βˆ€x,z:X,xβ‰ˆXzβ‡’(ev⁡U(i,x)↔ev⁡U(i,z)),Valid⁡(x)=βˆ€i,iβ€²:𝒫(U),iβ‰ˆπ’«iβ€²β‡’(iβ‰Όx↔iβ€²β‰Όx).\begin{aligned} i\approx_{\mathcal P}i' &=\forall x:X,\, \bigl(\operatorname{ev}_U(i,x)\leftrightarrow \operatorname{ev}_U(i',x)\bigr),\\ x\approx_X z &=\forall i:\mathcal P(U),\, \bigl(i\preceq x\leftrightarrow i\preceq z\bigr),\\ \operatorname{Ext}(i) &=\forall x,z:X,\, x\approx_X z\Rightarrow \bigl(\operatorname{ev}_U(i,x)\leftrightarrow \operatorname{ev}_U(i,z)\bigr),\\ \operatorname{Valid}(x) &=\forall i,i':\mathcal P(U),\, i\approx_{\mathcal P}i'\Rightarrow \bigl(i\preceq x\leftrightarrow i'\preceq x\bigr). \end{aligned}(5)

The probe double channel has the return path as its first path and rr as its second path.✦ For a double channel at a type VV, its predicate code type is the first path wrapper of the fourfold carrier, obtained by applying the double channel carrier four times to VV, and its parameter type is the second path wrapper of that predicate code type. The probe parameter type at AA is this parameter type at the probe raw type obtained by applying the initial path and then the raw path to AA.✦

For a double channel at a type VV, a source data bundle for observations d0,m,yd_0,m,y is a term built along its second path from the conditional whose guard and branches are their contravariant observations along its first path, in that order.✦

Bundle access observes its object along the second path against an access probe built along the first path.✦ The probe tests the supplied guard, returns logical truth when it holds, and otherwise composes the object observation callback with the supplied inner data callback.✦

The inner observation probe is built along the first path over the parameter type from the conditional that returns falsity on the dummy parameter test and otherwise returns the observation query at tag 77.✦

Constant falsity determines the bottom predicate code at VV along the first path over the fourfold carrier of VV.✦

Along the second path over the predicate code type, constant truth determines the dummy observation at VV.✦

For marker c0c_0 and payload cc, construct the encoded data point along the second path over the predicate code type. Its body conjoins the marker at tag 77 with the contravariant observation of cc after payload decoding.✦

To specialize a probe vv to a type AA, apply vv in data mode to the positive injection of AA along the lifting path.✦

The probe application observation at AA observes the specialized probe along the function path against the callback that applies the recovered function in data mode to the supplied single bundle and decodes the result through the tail.✦

For the channel relation with an admissibility guard, the equivalence class predicate of yy is the encoded predicate whose value at zz is that guarded observational relation between zz and yy.✦

The class union predicate of y,zy,z is the encoded predicate whose value at ww is the disjunction of the guarded observational relations from ww to yy and from ww to zz.✦

For a seven-query family, the class query at t∈{0,1,2}t\in\{0,1,2\} applies query t+1t+1 to the equivalence class predicate of the given element, with the supplied admissibility guard.✦

A false tagged edge from yy to zz requires class queries 00 and 11 at its endpoints and query 44 on their class union predicate. A true tagged edge instead uses endpoint classes 00 and 22 and query 55, throughout with the supplied admissibility guard.

Pair membership for a predicate pp and element aa at zz is the disjunction of the existence of an element satisfying pp with a false tagged edge to zz, and a true tagged edge from aa to zz, using the same queries and admissibility guard.✦

The pair predicate of pp and aa is the encoded predicate whose body is their pair membership formula.✦

At AA, encode a relational point yy as the encoded data point in the probe double channel at the probe raw type of AA, using the predicate path, the constant truth marker and the raw channel encoding of yy.✦

For m,ym,y at AA, take the source data bundle in the probe double channel at the probe parameter type of AA as their relational argument bundle. Its three observations are the dummy observation at the probe raw type of AA, mm and the relational point encoding of yy.

Observing a probe vv at AA against the relational argument bundle of m,ym,y by probe application gives its diamond observation against m,ym,y.✦

The relational base type of a type AA is the type obtained by applying the initial path to AA.✦ Encode the function assigning to yy the diamond observation of vv against m,ym,y as a predicate over this base type; this is the probe section predicate of vv at A,mA,m.✦

To obtain the relational pair predicate at A,mA,m, use the channel pair predicate over the relational base type of AA, taking roundtrip invariance conjoined with structural query 00 as its admissibility guard and the seven structural queries of mm as its query family.✦

Observation query 66 of mm against the relational pair predicate of the probe section predicate of vv and aa gives the section pair query of v,av,a at A,mA,m.✦

For v:𝒫(T)v:\mathcal P(T), observation invariance G(v)G(v), class representation Ct(v)C_t(v) and edge representation Et(v)E_t(v) are G(v)=βˆ€y,z:Y,AE⁡(y,z)β‡’(e(v,y)↔e(v,z)),Ct(v)=βˆƒx:X,βˆ€w:Y,(e(v,w)↔AE⁡(w,ΞΉt(x))),Et(v)=βˆƒx:X,βˆ€w:Y,(e(v,w)↔(AE⁡(w,ΞΉ0(x))∨AE⁡(w,ΞΉt(x)))).\begin{aligned} G(v)&=\forall y,z:Y,\, \operatorname{AE}(y,z)\Rightarrow \bigl(e(v,y)\leftrightarrow e(v,z)\bigr),\\ C_t(v)&=\exists x:X,\,\forall w:Y,\, \bigl(e(v,w)\leftrightarrow\operatorname{AE}(w,\iota_t(x))\bigr),\\ E_t(v)&=\exists x:X,\,\forall w:Y,\, \Bigl(e(v,w)\leftrightarrow \bigl(\operatorname{AE}(w,\iota_0(x))\vee \operatorname{AE}(w,\iota_t(x))\bigr)\Bigr). \end{aligned}(6) For v:𝒫(T)v:\mathcal P(T), the left copy predicate rv:𝒫(U)r_v:\mathcal P(U) is the encoded predicate Pred⁡U(x↦e(v,ΞΉ1(x)))\operatorname{Pred}_U(x\mapsto e(v,\iota_1(x))).✦ Relational closure R(v)R(v) of a predicate v:𝒫(T)v:\mathcal P(T) is the formula R(v)=Ext⁡(rv)β‡’βˆ€x:X,Valid⁡(x)β‡’(e(v,ΞΉ2(x))β‡’rvβ‰Όx).R(v)=\operatorname{Ext}(r_v)\Rightarrow \forall x:X,\, \operatorname{Valid}(x)\Rightarrow \bigl(e(v,\iota_2(x))\Rightarrow r_v\preceq x\bigr).(7)✦ The seven properties selected by the structural parameter are, in their order, (H0,H1,H2,H3,H4,H5,H6)=(G,C0,C1,C2,E1,E2,R).(H_0,H_1,H_2,H_3,H_4,H_5,H_6) =(G,C_0,C_1,C_2,E_1,E_2,R).(8)✦

The Chosen Parameter

The parameter m0m_0 is the structural query realization in the probe double channel at the raw type r(T)r(T). Its constant truth marker is built along rr from logical truth.✦ Its body is the finite marked disjunction of the seven branches HjH_j, each composed with the payload decoder and conjoined with the corresponding marker.✦ This realization uses the predicate path dd. For a predicate v:𝒫(T)v:\mathcal P(T) and one of the seven structural tags jj, the parameter query Qj(v)Q_j(v) is the formula obtained by tagged observation of m0m_0 against the payload encoding of vv at tag jj.✦

Roundtrip invariance Resp⁡T(v)\operatorname{Resp}_T(v) of a predicate v:𝒫(T)v:\mathcal P(T) is the formula requiring uniform equivalence of its evaluations before and after the roundtrip.✦ Goodness Good⁡(v)\operatorname{Good}(v) of a predicate v:𝒫(T)v:\mathcal P(T) and the specialized coded relation y∼zy\sim z between elements y,z:Yy,z:Y are the formulas Resp⁡T(v)=βˆ€y:Y,(e(v,y)↔e(v,rt⁡T(y))),Good⁡(v)=Resp⁡T(v)∧Q0(v),y∼z=βˆ€v:𝒫(T),Good⁡(v)β‡’(e(v,y)↔e(v,z)).\begin{aligned} \operatorname{Resp}_T(v) &=\forall y:Y,\, \bigl(e(v,y)\leftrightarrow e(v,\operatorname{rt}_T(y))\bigr),\\ \operatorname{Good}(v)&=\operatorname{Resp}_T(v)\wedge Q_0(v),\\ y\sim z&=\forall v:\mathcal P(T),\, \operatorname{Good}(v)\Rightarrow \bigl(e(v,y)\leftrightarrow e(v,z)\bigr). \end{aligned}(9) The admissibility guard in (9) is the specialized structural admissibility predicate at A0A_0 and m0m_0.

The Profile Conditions

The normal form sort profile πΔ(A)\pi_\Delta(A) of a term AA in context Ξ”\Delta is the set of all sorts typing its chosen normal form A#A^\#: πΔ(A)={s∈Sβˆ£Ξ”βŠ’A#:s},Ax⁡(s)={t∈S∣(s,t)βˆˆπ’œ},Out⁡(I,J)={cβˆ£βˆƒa∈I,βˆƒb∈J,(a,b,c)βˆˆβ„›}.\pi_\Delta(A)=\{s\in S\mid\Delta\vdash A^\#:s\},\qquad \operatorname{Ax}(s)=\{t\in S\mid(s,t)\in\mathcal A\},\qquad \operatorname{Out}(I,J)= \{c\mid\exists a\in I,\exists b\in J,\ (a,b,c)\in\mathcal R\}.(10) A feasible profile IβŠ†SI\subseteq S is a nonempty set I=πΔ(A)I=\pi_\Delta(A) for some normal term AA in a valid context Ξ”\Delta.✦ A valid context is a context obtained from the empty context by extensions Ξ”,x:B\Delta,x:B with Ξ”βŠ’B:s\Delta\vdash B:s for some sort ss.✦ A feasible product profile triple (I,J,K)(I,J,K) consists of feasible profiles with K=Out⁡(I,J)K=\operatorname{Out}(I,J), using (10).✦ Normality here is for full Ξ²\beta reduction, including annotations and binder bodies. Relative to the collection CC of profiles used by LL, a terminal profile JJ is a feasible profile occurring in a feasible product profile triple (D,J,K)(D,J,K) with D,K∈CD,K\in C.✦ The product output profile Out⁡(I,J)\operatorname{Out}(I,J) consists of every output sort permitted by a product rule whose domain sort belongs to II and whose codomain sort belongs to JJ.✦

In the primary signed graph, a positive inclusion edge goes from a feasible profile II to a feasible proper subset JJ. Every feasible product profile triple (I,J,K)(I,J,K) supplies a negative domain edge from II to KK and a positive codomain edge from JJ to KK, and these are all the primary edges. Positive reachability means a finite, possibly empty path all of whose edges are positive.✦ Relative to a collection CC of profiles, an observation layer II is a profile positively reachable in the primary signed graph from a nonempty axiom profile Ax⁡(s)\operatorname{Ax}(s) with {s}∈C\{s\}\in C.✦ Relative to a collection CC of profiles, a direct observation layer II is an observation layer for which a feasible product profile triple (I,J,K)(I,J,K) has J,K∈CJ,K\in C.✦

The ambient logical calculus has fixed feasible witnesses whose types are data-only normal forms sorted at every sort in the assigned profile, with a chosen inhabitant of each type; both types and inhabitants have data mode.✦ For every two terminal profiles, the calculus has a uniform typed open proof transformer between their chosen witness types. It fixes a feasible product profile triple (dL,bL,kL)(d_L,b_L,k_L) with all three profiles in CC, an even incoming proof path from bLb_L to dLd_L, and an even outgoing proof path from kLk_L to bLb_L. For every direct observation layer DD, a quantifier route chooses a feasible product profile triple (D,JD,KD)(D,J_D,K_D), an even incoming proof path from bLb_L to JDJ_D, and an even outgoing proof path from KDK_D to bLb_L. The probe has feasible product profile triples (I,J,k)(I,J,k) and (p,n,g)(p,n,g), a seed ss with nonempty axiom profile, and paths with endpoints Ξ±:{s}β†’k,r:kβ†’p,back:pβ†’k,fun:gβ†’J,lift:Ax⁡(s)β†’I.\alpha:\{s\}\longrightarrow k,\quad r:k\longrightarrow p,\quad \mathrm{back}:p\longrightarrow k,\quad \mathrm{fun}:g\longrightarrow J,\quad \mathrm{lift}:\operatorname{Ax}(s)\longrightarrow I.(11) In (11), the return and raw paths are odd, the function path is even, the lifting path is positive, and the data tail ends at nn. Let jj and hh be the target profiles of the logical and predicate channels. The shared logical channel runs from kk to jj, and the predicate channel from pp to hh. The inner observation double channel runs from pp through kk to pp, and the outer one from pp through kk to jj. The type A0A_0 satisfies Ξ“βŠ’A0:s\Gamma\vdash A_0:s in data mode. In addition to the carriers of (2), the specialization requires Y=β„“(T),Oout,2=β„“,Oout,1(Oin(r(U)))=T.Y=\ell(T),\qquad O_{\mathrm{out},2}=\ell,\qquad O_{\mathrm{out},1}\bigl(O_{\mathrm{in}}(r(U))\bigr)=T.(12) The subscripts in (12) designate the first and second paths of the outer double channel. Each negative data tail has an even decoding proof path containing a negative step.

The Claim

Theorem 1 (Lemma 49). Let the specialized relational coding and the logical quantifiers satisfy their path, sorting and carrier conditions. For every vv with Ξ“βŠ’v:𝒫(T)\Gamma\vdash v:\mathcal P(T) and data mode under Ξ³\gamma, Good⁡(v)≃Γ,Ξ³G(v).\operatorname{Good}(v)\simeq_{\Gamma,\gamma}G(v).(13)✦ For y,zy,z with Ξ“βŠ’y:Y\Gamma\vdash y:Y and Ξ“βŠ’z:Y\Gamma\vdash z:Y, both in data mode, (y∼z)≃Γ,Ξ³AE⁡(y,z).(y\sim z)\simeq_{\Gamma,\gamma}\operatorname{AE}(y,z).(14)✦ For x,zx,z with Ξ“βŠ’x:X\Gamma\vdash x:X and Ξ“βŠ’z:X\Gamma\vdash z:X, both in data mode, and every tag t∈{0,1,2}t\in\{0,1,2\}, (ΞΉt(x)∼ιt(z))≃Γ,Ξ³(xβ‰ˆXz).\bigl(\iota_t(x)\sim\iota_t(z)\bigr) \simeq_{\Gamma,\gamma}(x\approx_X z).(15)✦ For such x,zx,z and distinct tags t,u∈{0,1,2}t,u\in\{0,1,2\}, Ξ“βŠ©Ξ³Β¬(ΞΉt(x)∼ιu(z)).\Gamma\Vdash_\gamma\neg\bigl(\iota_t(x)\sim\iota_u(z)\bigr).(16)✦

What Is Checked

The statements (13), (14), (15) and (16) concern the specialized relational coding under these profile and shape conditions, with labelled typing and the data and proof mode discipline interpreted by (1).