A one-variable main conjecture for CM elliptic curves compares a torsion arithmetic Iwasawa module with a measure interpolating special Hecke L-function values. The classical bounded-measure formulation uses a prime of good reduction which splits in the CM field, hence gives ordinary reduction. We first describe the one-prime division-tower formulation, including the analytic construction, and then explain its cyclotomic Selmer counterpart and the change at supersingular primes.
For a precise basic setting, take with complex multiplication by , of class number one, and a split prime of good reduction, prime to the Hecke conductor . Write , choose the -adic embedding corresponding to , and let . Its formal group of an elliptic curve has height one, with . The one-prime division tower
has of order prime to and . Its CM character describes the action on the prime-primary Tate module. Let and . We extend coefficients to the complete unramified ring to include character values and a CM period.
Let be the Galois group over of the maximal abelian pro- extension unramified outside primes above . Let be the analogous everywhere-unramified group, equivalently the norm inverse limit of the -parts of the ideal class groups. Write for the norm limit of principal local units at , for the closures of global units in , and for the closures of elliptic units. Roots of unity and finite initial-layer errors are retained when formulating integral statements; after passage to a finite-character component they do not change a characteristic ideal.
The class-field unit sequence, obtained by taking norm limits in global Artin reciprocity, gives
In this ordinary one-prime tower, the unit and global-unit terms have the same Iwasawa rank, and the displayed quotients and are torsion on their character components. The equality
is the global-units/class-groups form of the CM main theorem. Multiplicativity of characteristic ideals in the sequence then gives
The analytic construction must identify the latter ideal with a genuinely interpolating p-adic L-function; it is not enough to rename an arbitrary characteristic power series as an L-function.
Here is that construction. For an auxiliary ideal prime to , a suitably normalized rational elliptic function has divisor a multiple of
Translate by a generator of the conductor torsion and take the product over its ray-class orbit. This produces a rational function defined over which is a unit series at the identity. Evaluating it on compatible -division points gives elliptic units. The reason for the norm relations is geometric: multiplication by groups the translated zeros and poles into fibers, so the product over has the divisor of . The normalization fixes the constant, giving the local identity
This is also the norm relation of the Coleman power series for a CM formal group interpolating the resulting local units.
An integral logarithmic correction makes the passage from those units to a bounded measure explicit. In the chosen split completion , put
Since and the coefficients of reduce to , the ratio is in . For , its logarithm divided by is integral. Constants are treated using the usual -adic logarithm on units. This proves the integrality needed for the Frobenius-corrected logarithm; applying an uncorrected logarithm alone would not prove boundedness.
Choose an isomorphism over , and normalize the CM period by
Then is the Amice transform of a unique -valued p-adic measure on . The norm relation above gives
the first logarithmic sum is and the second is the same sum, because and there are points. On the measure side the sum selects and multiplies by , so this identity says that the measure is supported on . Transport it to by the CM character and project to the desired component.
The moments can now be calculated rather than postulated. With ,
The factor appears because is multiplication by in the additive formal-logarithm coordinate. The remaining derivative is a Coates–Wiles homomorphism. The complex product for the elliptic function makes it a sum of shifted inverse powers of lattice elements. Grouping these lattice elements by ray-class ideals identifies that sum with , multiplied by the auxiliary factor . Here omits primes dividing the original conductor, including when the conductor of the power character becomes smaller.
Normalize the rational functions and matching complex and -adic periods consistently, and remove the auxiliary factors by the distribution relations. They are evaluations of in the Iwasawa algebra. On each finite-character component an auxiliary ideal can be chosen making this element a unit: its residues at and can be prescribed independently by the Chinese remainder theorem. The resulting integral measure satisfies, on its branch and with in the corresponding residue class,
The exact normalization of periods matters for this equality; multiplication by an integral unit does not affect the ideal in the main conjecture. Finite-order twists are obtained by inserting a character in the ray-class sums, with its local Gauss sum. The lattice calculation proves algebraicity of the normalized complex values, while the integral series construction proves their -adic interpolation. These are different steps.
The Coleman measure map identifies with a rank-one Iwasawa algebra lattice, up to finite errors, and takes to the ideal generated by the corresponding power series . The conjecture therefore has the concrete form
Its proof uses the Euler system of elliptic units. Adjoining auxiliary ray moduli supplies norm relations with factors . Derivative operators in the auxiliary cyclic groups produce Kolyvagin classes; their controlled valuations relate a new auxiliary prime to its ideal class. Chebotarev density theorem permits the required Frobenius choices. Successively choosing these primes bounds the class-group elementary divisors by the elliptic-unit index, giving the difficult characteristic-ideal divisibility. The opposite comparison comes from the class-number/unit-index formula and passage through the tower. Their equality, combined with the exact sequence above and the logarithmic computation, is the proof mechanism. The ordinary split-prime setting of this formulation is specified in Rubin's one-variable main-conjecture paper.
For completeness, the familiar cyclotomic formulation concerns a different tower. If has complex multiplication and good ordinary reduction at , put and
For a finite layer , its Selmer group of an elliptic curve is
where contains the bad primes and primes above ; at infinite level take the direct limit. Thus it imposes the local Kummer conditions and sits between rational points tensored with and the -primary Tate–Shafarevich group. Its dual is a torsion Iwasawa module, and the cyclotomic main conjecture is
With compatible reciprocity conventions, the ordinary function has the finite-character interpolation
where is the -adic unit root of and ; characters of this cyclotomic -extension are even. The two-variable CM measure specializes to this line after the CM-character twist. Kummer theory, class-field descriptions of the prime-primary local conditions and control of finite-layer errors identify its algebraic counterpart with the displayed Selmer dual. The one-prime division tower is not itself the cyclotomic tower, so this requires the twist and specialization argument, not just changing its name.
Finally, at a good inert prime the curve has supersingular reduction. The ordinary unit root is absent, the unmodified cyclotomic Selmer dual is not torsion, and the ordinary bounded-function argument does not apply. For CM curves over and odd supersingular , signed local trace conditions and signed functions give instead , as proved in Pollack and Rubin's supersingular main-conjecture theorem. Thus the ordinary hypotheses and the analytic local correction are essential parts of the essay's main-conjecture statement.
The construction must first specify the field of definition of the complex multiplication. Let be a CM elliptic curve and let be its geometric CM field, embedded by its action on an invariant differential. In characteristic zero this action is injective. Consequently every geometric endomorphism is defined over : a Galois automorphism fixing fixes its differential scalar and therefore fixes the endomorphism itself. Initially assume already contains ; we return to the other case below.
Choose a complex embedding and an analytic complex uniformization of an elliptic curve
where is a lattice in , stable under the CM order . It need not be the maximal order or a principal ideal. Its torsion is described by the finite adelic quotients of this lattice.
The input from the CM reciprocity theorem is precise: if is an idele of , , and , then takes the CM torus to the torus with lattice ; on torsion the comparison is induced by multiplication by . This reciprocity statement comes from the ideal action on CM lattices: an ideal gives the isogeny to , composition multiplies ideals, and adding torsion level structure removes automorphism ambiguity. The resulting action on algebraic CM points is the ray-class Artin action. This is the deep CM input, not an assertion that an arbitrary abelian Galois representation automatically comes from an algebraic Hecke character.
Because fixes , the algebraic curve is . Comparing the reciprocity uniformization with the original one therefore gives multiplication by a scalar
The scalar belongs to : a complex scalar carrying one lattice contained in to another is the quotient of two nonzero lattice elements. Its torsion action is
The uniqueness of the torsion-compatible comparison is important: two possible scalars would induce the same action on all torsion, hence their difference would annihilate torsion of arbitrarily large order and be zero. Multiplying the torsion-action formulas for and now proves .
For , Artin reciprocity sends its principal idele to the identity. The archimedean factors are complex and connected, so its finite part also has trivial Artin image. The preceding formula therefore gives
It follows that
is trivial on principal ideles and defines a character of the idèle class group of . This establishes the multiplicativity and principal-idele condition, rather than merely prescribing Frobenius values and hoping that they extend.
We must still prove continuity and a finite conductor. For a finite place and a prime different from its residue characteristic, an idele supported on the units at has trivial -component of its norm. Thus its action on the Tate module is exactly multiplication by . At a place of good reduction, inertia acts trivially, so . At a bad place, a CM curve has potentially good reduction, and inertia on the prime-to-residue-characteristic Tate module has finite image. Hence is trivial on an open subgroup of the local units there. There are only finitely many bad places. Together these facts give an open finite-congruence kernel on the unit ideles and prove continuity. The archimedean norm factor is already continuous. The potentially-good-reduction and inertia facts are the reduction-theoretic inputs in Serre and Tate's treatment of CM abelian varieties.
Consequently is a Hecke character. In ideal notation, for ,
so its infinity type of a Hecke character is the type determined by the chosen CM embedding, with the inverse archimedean norm in the idele convention. Nonmaximal orders merely contribute additional finite-level conditions; the proof retains their lattice and does not replace it by the maximal-order lattice without justification.
At a good prime , choose a uniformizer idele and a prime different from the residue characteristic. Its norm has trivial -component, so arithmetic Frobenius acts on the Tate module by . Specialization of prime-to-residue-characteristic torsion identifies this with the reduced Frobenius isogeny. Thus
The corresponding Hasse-Weil L-function factors as , including the local factors determined by inertia. This explains why the Hecke character encodes both the torsion representations and the Euler factors.
If does not contain , apply the construction over . The two conjugate CM characters are exchanged by the nontrivial automorphism of , and the two-dimensional Galois representation over is their induced representation. In that case the natural description is by induction, rather than a pair of one-dimensional characters over . Question 3 supplies a concrete example: does not define , whereas does. This completes the existence argument for an arbitrary CM curve over a number field, with its field of definition correctly accounted for.
The map
is an automorphism of the elliptic curve over , fixes its identity point, and squares to . For the invariant differential on an elliptic curve , it satisfies . Thus , with the indicated embedding of .
Under analytic complex uniformization of an elliptic curve , this automorphism acts as multiplication by , so is stable under . It is a rank-one torsion-free module over the Gaussian integers. Since that ring is a principal ideal domain, . Every analytic endomorphism is multiplication by an with , which here says exactly . Therefore the geometric endomorphism ring of an elliptic curve is , and all its elements are generated over by the displayed -defined automorphism. Thus .
If an endomorphism is defined over , its pullback on the rational differential has scalar in . The differential action is injective in characteristic zero, because a nonzero isogeny of elliptic curves is separable and has nonzero differential. Its scalar must therefore lie in . Conversely integer multiplication is defined over . Hence
Put . The valuations of the differences of the three nonidentity fourth roots of unity from are
For instance , and the other two differences are associates of . None is divisible by . The group has elements, so the four units map bijectively to it. Every ideal prime to consequently has a unique primary Gaussian integer generator.
The Grössencharacter of the square-lattice elliptic curve is explicitly
For integral ideals prime to , the generator is an odd Gaussian integer; extend multiplicatively to fractional ideals prime to . Products of primary generators are primary, so this is a Hecke character with ideal infinity type : on a principal ideal with generator , its value is . Its conductor divides , and cannot divide , since would force , whereas the unit ideal has value . Thus the displayed conductor is exact.
We next verify directly that this character gives the Frobenius values of this particular curve, fixing the unit ambiguity. The point
satisfies, by the chord-and-tangent addition law,
For the first addition, and the chord has slope . Since is killed by but not by , it generates the eight-element -module . Thus all this torsion is rational over .
The model has discriminant , so it has good reduction at every . First suppose lies over a split rational prime . Then its residue field is , , and has norm . Reduction of still has degree and has zero differential, since its differential scalar belongs to . It is therefore purely inseparable of degree . Factoring out the Frobenius isogeny gives
for an automorphism of the reduction. In characteristic at least five, the automorphisms of this model fixing are the four Gaussian units: comparing coefficients in gives . Prime-to- torsion reduces injectively. Both and fix the reduction of , the former by the primary congruence and the latter by its rationality over the residue field. Thus fixes this entire module. The unit congruences proved above force .
For an inert rational prime , we have , , and its primary generator is , because implies . To determine Frobenius without excluding , write
Replacing by changes each nonzero quadratic-character value to its negative, because is a nonsquare. Hence . The Frobenius characteristic equation gives . This proves the asserted Frobenius value at inert primes as well, and identifies the displayed Hecke character with that of .
Finally, the formal parameter is defined over the local ring at . Since the reduced endomorphism is the -power Frobenius isogeny, with , its pullback sends to . Passing to the completion at the identity gives the equality of formal series
The coefficient ring is complete for its -adic norm, with uniformizer and residue field . Its residue field is infinite, so is not compact. The measure construction uses completeness, not compactness of the coefficient ring.
An -valued p-adic measure on is a bounded -linear functional
on the continuous functions with the uniform norm. Equivalently it assigns a value to each clopen subset, finitely additively on disjoint unions. Integration of a locally constant function is defined by
Refining finite clopen partitions shows that this is independent of the expression. Since every , the nonarchimedean triangle inequality gives .
Locally constant functions are uniformly dense in . Indeed, for every , continuity and compactness of supply a finite clopen partition on which is constant modulo ; choose one value of on each part. These approximations converge uniformly, and completeness of extends integration uniquely. Conversely, a bounded functional gives and finite additivity. This proves the equivalence of the two definitions. Ordinary countable additivity of real measures is not part of the definition.
Define the Iwasawa algebra by
where runs over open subgroups. Since is an abelian profinite group, every such quotient is a finite abelian group. The transition map for sends a basis element to , so sums coefficients over the fibers.
A p-adic measure defines an element of this inverse limit by
Finite additivity is precisely compatibility under the transition maps. Conversely, compatible coefficients define values on every coset and hence on every clopen subset: any clopen subset is a union of cosets of one open subgroup, by compactness and a finite common refinement. Compatibility makes the resulting value independent of that subgroup. The preceding integration argument then constructs the p-adic measure.
Finally, multiplication is convolution of p-adic measures. On a finite quotient its coefficients are
This is exactly the coefficient of in . Equivalently,
first for locally constant , then by uniform approximation. The Dirac measure at the identity corresponds to the identity element of the group algebra. We have proved a bijection respecting addition, scalar multiplication and multiplication:
Fix a separable closure . The Tate module and rational Tate module are
where the transition maps are multiplication by . Since is different from the characteristic, and these transition maps are surjective. Choosing compatible bases gives , and . Both spaces carry the natural continuous action of the absolute Galois group of and of the base-defined endomorphism ring of an elliptic curve.
For (i), choose a nonrational . Such an element exists because the intersection of the given two-dimensional rational subalgebra with the integral endomorphism ring contains a full lattice. The characteristic equation of an elliptic endomorphism is
Since is nonrational, its minimal polynomial over has degree two. Comparing the two monic equations gives and . The same is the characteristic polynomial on the rational Tate module: the Weil pairing gives determinant , and applying it to gives trace . Thus on the characteristic polynomial is
If is a quadratic field over , a unital -action on this two-dimensional -space makes it a one-dimensional -space. If , the two roots above are distinct in and their eigenspaces each have dimension one. The two idempotent factors therefore have equal rank one, so the module is again free of rank one. This also proves faithfulness in the split case. Hence
For (ii), first prove an integral divisibility statement. If maps into , then its induced action on is zero. The kernel of an isogeny of is therefore contained in the kernel of . The quotient property of this isogeny of elliptic curves gives a unique endomorphism with
This factorization is defined over : applying any descent automorphism preserves it, and uniqueness fixes . Inside the rational endomorphism ring, belongs to , so . Consequently
Here the prime-to-characteristic torsion is étale, so vanishing on its geometric points is vanishing on the entire kernel scheme.
Every element of preserves the Tate module by continuity. Conversely suppose preserves it. Choose with , and approximate it by modulo . Write with . Then , so the divisibility statement gives . It follows that . Thus the CM Tate module stabilizer is
This does not presume freeness of the integral Tate module over a nonmaximal order.
For (iii), every element of is defined over , so the Galois representation commutes with and therefore with . By (i), its action on is multiplication by an element of . The action and its inverse preserve ; by (ii), this scalar and its inverse belong to . Hence
A regular category has finite limits, coequalizers of kernel pairs, and regular epimorphisms stable under pullback. Equivalently, every morphism factors as a regular epimorphism followed by a monomorphism, and these factorizations are stable under pullback. A regular functor preserves finite limits and regular epimorphisms, hence also preserves these image factorizations.
For this question, the capital regular category convention is the one in which every well-supported object is a well-pointed object in a category. Here an object is well-supported if is a regular epimorphism. It is well-pointed if a subobject through which every point factors must be the whole object. This convention is documented in Definitions 1.1–1.3. It imposes no such point-detection requirement on objects whose support is a proper subobject of . In particular, it should not be replaced here by the stronger assertion that reflects every isomorphism.
First prove two facts about a capital regular category . Every well-supported object has a point. Otherwise is well-supported and has no points, so its diagonal contains all points and is invertible. That makes a subterminal object; the regular epimorphism is then also monic and hence invertible, contradicting the absence of a point. If is a regular epimorphism and is a point, its pullback in a category is a well-supported object over , so it has a point lifting . Thus the representable functor
preserves regular epimorphisms; it already preserves all existing limits. Hence is a regular functor. Second, if is a singleton, its unique point is a monomorphism, is well-supported, and contains all points of . Well-pointedness makes invertible. Thus reflects terminal objects, even though it need not reflect arbitrary isomorphisms.
Now let be a small category that is a regular category. For each object , its slice category is small and regular: finite limits, images and the stability of regular epimorphisms are computed using the corresponding constructions in . By the permitted capitalization result choose a conservative functor that is regular,
with a small capital regular category. The base-change functor , given by , preserves finite limits and regular epimorphisms. So does
Taking these as the coordinates defines a regular functor
The exponent is a set of indices, so this is a power of the Category of sets, with finite limits and regular epimorphisms computed coordinatewise.
To prove reflection of isomorphisms, suppose is invertible for . In , the object is the pullback of along the point given by the diagonal . Applying shows that is the fibre of the bijection over this diagonal point, and is a singleton. The preceding lemma gives through its canonical arrow to . Since reflects isomorphisms, is invertible. Its underlying morphism is . Therefore
In the small-category setting of this representation result, the elementary condition for replacing the power by one copy of is:
This is almost total support for a regular category. “Strict initial” means initial and every arrow into it is invertible; the condition allows the case in which all objects are well-supported and no initial object exists.
For necessity, let be regular and conservative. If , factor through its support of an object in a regular category . Preservation of images gives . Conservativity makes invertible, so is well-supported. If , then for every the projection is sent to the bijection , hence is invertible. Its inverse followed by the other projection supplies an arrow . For any two such arrows, their equalizer is sent to a bijection onto , hence is invertible, proving uniqueness. Thus is initial. If exists, then exists, so and that arrow is sent to a bijection. Conservativity makes it invertible. Hence is strict initial.
For sufficiency, choose the permitted conservative regular functor into a small capital regular category, and put . This is regular. A well-supported object of is sent to a well-supported object of , which has a point. A non-well-supported object is strict initial by the condition, and is consequently a proper subterminal object. The functor preserves subterminality and reflects the noninvertibility of its arrow to ; its image is a proper subterminal object and has no points. Thus is empty exactly on the non-well-supported objects.
Suppose is a bijection for . If both sets are empty, and are strict initial objects and is invertible. Otherwise are well-supported. Factor as a regular epimorphism followed by a monomorphism . Surjectivity of makes every point of factor through ; capitality of makes invertible. Hence is a regular epimorphism. Its kernel pair is well-supported, because and are regular epimorphisms. Injectivity of implies that every point of has equal coordinates, and therefore factors through the diagonal . Since is well-pointed, this diagonal is invertible. Thus is also monic, hence an isomorphism, and conservativity of makes invertible. This proves sufficiency with actual point and image arguments, rather than asserting that global sections of every capital regular category is conservative.
Let be a final functor, and let be a cocone under a diagram . For , choose an object of the nonempty comma category and define
If is a morphism in that comma category, then , so the cocone equation gives . Equality also holds along a reversed arrow, and hence along any zigzag. Connectedness makes the definition independent of the chosen object. For , use and to obtain . Thus is a cocone under . Taking shows that it extends .
Every extension must satisfy by its cocone equation, proving uniqueness. The construction commutes with postcomposition by a morphism . Hence restriction gives a natural bijection between the sets of cocones of and with any fixed vertex. A universal cocone for therefore extends to a universal cocone for . In particular,
whenever the latter exists, and existence of all colimits of shape implies existence of all colimits of shape .
For the factorization of an arbitrary functor , let
the set of connected components of a category, where components are defined by finite zigzags. Precomposition by gives a functor and therefore a map . Identity and composition laws make a categorical presheaf. This is a set-valued construction when the relevant comma categories are small or have only a set of components, as in the usual small-category convention for this factorization.
Take to be the category of elements of this presheaf. Its objects are pairs with ; a morphism is an arrow satisfying . Define by forgetting . Given and an object , the unique lift into that object has source and underlying arrow . This proves that is a discrete fibration.
Define . For , take . The arrow in joins to , so the required component equation holds. It follows that is a functor and .
Finally, an object of is a pair whose component in is . Its morphisms are exactly the morphisms of that comma category between objects in . Consequently is precisely the full connected component , and is nonempty and connected. Thus is final, and every such functor factors as a final functor followed by a discrete fibration.
A balanced category is one in which every morphism that is both a monomorphism and an epimorphism is an isomorphism. A strong monomorphism is a monomorphism right orthogonal to all epimorphisms: whenever is epic and , satisfy , there is a unique with and . A regular monomorphism is an equalizer of some parallel pair of morphisms. Its equalizer property supplies the indicated lift after epic cancellation, so every regular monomorphism is strong.
If every monomorphism is strong and is also epic, apply strongness to the square whose left and right sides are both , with horizontal arrows and . The lift satisfies and . Thus is invertible and the category is balanced.
Conversely, suppose the category is balanced and has pullbacks in a category. Given the lifting square above, form with projections , . The pair gives with , . The pullback of the monomorphism is monic, so is monic. It is also epic: implies , hence because is epic. Balancedness makes invertible. Now satisfies both required equations, and monicity of proves uniqueness. Hence in a balanced category with pullbacks, every monomorphism is strong.
For the finite example, choose a strong monomorphism that is not regular. It is not an isomorphism, since an isomorphism is the equalizer of an identical pair. It is not epic either, since an epic strong monomorphism is invertible by the first argument. Consequently there are distinct arrows with . Since is not the equalizer of this pair, there is a map with that does not factor through . Monicity rules out failure of uniqueness, so failure of existence is the only possible failure of its equalizer property.
Define to have four formally distinct objects and exactly the following six nonidentity arrows:
Their only nontrivial composites are
There are no longer composable strings of nonidentity arrows, so these rules define an associative category. Sending to , respectively, and the six arrows to , defines a functor. It is a faithful functor: every hom-set has at most one arrow except , and their images are distinct. Faithfulness does not require the four image objects to be distinct.
The map is monic, by direct inspection of arrows into . Neither nor is epic, since the distinct pair agrees after either. Every other nonidentity arrow has target and is epic, because there is only one arrow out of . A lifting square against whose left side is one of these epimorphisms would require an arrow , which does not exist. Squares whose left side is an identity have the evident unique lift. Thus is a strong monomorphism.
It is not a regular monomorphism. The only distinct parallel pair out of is , and equalizes it but has no factorization through , since there is no arrow . An identical pair has as its equalizer and cannot have the noninvertible as its equalizer. Therefore
The Yoneda embedding is an exact representation of a category, but does not by itself reduce the subject to the study of arbitrary functor categories. There are both a size issue and a structure issue.
Let be a locally small category and set . For a categorical presheaf , a natural transformation is determined by . Indeed, naturality along forces
Conversely this formula defines a natural transformation for each , since functoriality gives naturality along every . Taking gives
Thus the Yoneda embedding is full and faithful, and is equivalent to its full subcategory of representable presheaves. The indexing category in the usual version is , not necessarily a small category chosen independently of .
Local smallness makes the values sets; it does not make the collection of objects a set. For a large indexing category, the collection of natural transformations between arbitrary set-valued functors can itself be a proper class. For example, take the discrete category on the class of ordinals and the constant two-element functor. Its endomorphisms include an independently chosen function at each ordinal. A small functor category cannot encode this collection without changing universes. The Yoneda lemma remains meaningful when the source is representable because the displayed bijection identifies its natural transformations with a set. It does not make the entire large functor category locally small.
Even for a small category, the representable image is usually a very special part of its presheaf category. If is the terminal category, its presheaf category is , but its representable image consists only of a singleton. The presheaf category has all small categorical limits and colimits, whereas an arbitrary full subcategory need have almost none. The Yoneda embedding preserves existing limits because representable functors preserve limits. It generally does not preserve colimits. For example, the coproduct of two singleton sets in category of finite sets is a two-element set, but evaluating the proposed presheaf coproduct at a two-element set gives
The pointwise coproduct therefore leaves the representable image. Identifying which presheaves are representable, and which ambient constructions return to the original category, is precisely substantive category theory.
The representation is nonetheless powerful. The Yoneda lemma replaces a morphism by all its probes, converts a universal property into a representation of a functor, and makes the presheaf category a natural place to add formal colimits. A small presheaf category is the free free cocompletion under small colimits: every presheaf is a colimit of representables indexed by its category of elements. But this is a controlled enlargement of the category, not an identification of its original objects with all functors. The stronger conclusion in the proposed assertion therefore does not follow from the full-faithfulness statement.
The general adjoint functor theorem says that, for a functor between locally small categories, with a complete category, has a left adjoint if and only if it preserves small limits and satisfies the solution-set condition. The solution-set condition means that for each there is a set of arrows such that every factors as for some and some .
Here is the smallness argument behind the theorem. A complete category that is locally small and has a weakly initial set has an initial object. Form the product in a category . It is weakly initial, since a projection followed by a chosen arrow gives an arrow . Take the simultaneous equalizer of all endomorphisms of with ; the family is a set by local smallness. The object is also weakly initial. For parallel arrows , take their equalizer and a map . Since is an endomorphism of , the defining property of gives . Cancellation of the monomorphism gives . Thus the monomorphism has a right inverse and is an isomorphism, so . Existence of arrows out of was already established, proving the initial-object lemma for complete categories with a weakly initial set.
If preserves small categorical limits, the comma category is complete: take the limiting object in , and use preservation to assemble the arrows from . It is locally small, and the solution-set condition provides a weakly initial set. Its initial object is an arrow with a natural bijection
For a morphism , define by . Uniqueness proves functoriality and naturality, giving a left adjoint . Conversely, a right adjoint preserves categorical limits, because applying the adjunction converts a limiting cone over a diagram into the corresponding limiting cone of hom-sets. Its adjunction unit at is a singleton solution set. This proves both directions of the general adjoint functor theorem.
Apply it to the inclusion , where is the full subcategory of compact Hausdorff spaces. A small product in a category of compact Hausdorff spaces is again compact Hausdorff by the Tychonoff theorem. An equalizer of two continuous functions into a Hausdorff space is closed, hence compact Hausdorff. Products and equalizers construct all small categorical limits, and the inclusion preserves these constructions. Both categories are locally small.
Fix a topological space and a continuous function into a compact Hausdorff space. The closure is compact Hausdorff and has a dense subset of cardinality at most . The permitted cardinal estimate gives
For , use . Choose a set of representatives of all compact Hausdorff topologies on underlying sets of cardinality at most this bound, and all continuous functions from to those representatives. Every factors through one of them, using the inclusion . This is the solution-set condition. Consequently the inclusion of compact Hausdorff spaces has a left adjoint, the compact Hausdorff reflection:
The universal arrow has dense image: factoring it through its closed image and applying uniqueness supplies an inverse to that image inclusion. It need not be an embedding when is not sufficiently separated; the assertion is a compact Hausdorff reflection, rather than a claim that every topological space embeds in a compact Hausdorff space.
For completeness, the alternative Special adjoint functor theorem can be proved from the same initial-object argument. Its limit form assumes that is complete, locally small and well-powered, with a small cogenerating family ; then , with locally small, has a left adjoint precisely when it preserves small limits. To prove sufficiency, start with and intersect all subobjects through which factors after applying . Well-poweredness makes this a small intersection of subobjects, and preservation of limits gives a smallest supporting object . For maps , equality makes their equalizer another supporting subobject. Minimality forces that equalizer to be invertible, so . Therefore
is an injective function. The evaluation embedding into cogenerator products embeds in , where is the realized subset of the fixed set . There are only a set of choices of these subsets, only a set of subobjects of each resulting product, and only a set of arrows from to each -image. They give a solution set, so the already proved general adjoint functor theorem applies. Necessity is again preservation of limits by a right adjoint. For , monomorphisms are embeddings onto closed subspaces, and is a coseparator because continuous functions to it separate points. Thus the Special adjoint functor theorem gives the same reflection directly.
A model-theoretic ultralimit is the direct limit of a chain of successive ultrapowers along their diagonal elementary maps. It is not the topological notion of a limit along an ultrafilter. We first prove the precise relative embedding lemma needed to synchronize two such limits.
If , then embeds elementarily into an ultrapower of , with the embedding restricting on to its diagonal map. To prove this Frayne embedding lemma, use the elementary diagram of with the elements of treated as named parameters. Every finite fragment of that diagram has a realization in : replace its finitely many names from by variables and existentially quantify their conjunction. It is true in , and hence true in by elementarity. Choose these simultaneous realizations for each finite fragment, keeping every parameter from interpreted as itself; give unmentioned names an arbitrary default in the nonempty .
On the index set of finite fragments, extend the filter of cones to an ultrafilter . For each , let be its value in the chosen coordinate realization. Every diagram sentence is satisfied on its cone, so the Łoś theorem makes an elementary embedding . Inequalities in the diagram ensure injectivity. If , all its coordinates equal , giving exactly the diagonal map.
There is also the initial version: if merely , then embeds elementarily into an ultrapower of . A finite fragment of the elementary diagram of , with all its names replaced by variables, yields an existential sentence true in and hence in . The identical finite-fragment and cone construction therefore proves this version without already assuming an embedding.
Start with . Use the initial version to place elementarily in an ultrapower of . Relabel embedded copies so we can regard the elementary maps as inclusions. Apply the relative version to to obtain an ultrapower of containing , and whose inclusion of is its diagonal map. Next apply it to to obtain an ultrapower of containing , again fixing the earlier diagonal inclusion. Continuing gives an elementary chain
The common union is elementary over every stage by the elementary chain theorem, and
The first presentation is a model-theoretic ultralimit of . The second, preceded by the initial diagonal embedding of into , is a model-theoretic ultralimit of . Each consecutive same-letter inclusion is precisely the corresponding ultrapower diagonal map, not just an unspecified elementary embedding. Thus the two limits are identified with the same structure , proving
Let be a full Skolem expansion of an infinite structure , choose distinct for , and extend the cofinite filter to a nonprincipal ultrafilter on . For a finite subset of the desired index order , define the ordered Fubini product of ultrafilters on by
where means . These nested tests decide complements and preserve intersections by induction on , so is indeed an ultrafilter. For the empty set use the one-point index set and its principal ultrafilter.
Put . If , projection of coordinates induces
A truth test independent of a coordinate is unchanged by its dummy ultrafilter quantifier. Consequently pullback preserves exactly the -large subsets, proving that the map is well defined and injective. The Łoś theorem, applied to every formula rather than just to equality, proves that it is an elementary embedding. Projections compose, so these elementary maps form a coherent directed system.
Take its directed limit of elementary embeddings . Explicitly identify two stage elements when their images agree at a common larger stage, and interpret functions and relations there. For an existential formula true in the limit, its parameters and a witness occur at a common stage; elementarity then pulls truth back to the stage of the original parameters. This verifies that all stage maps into the limit are elementary.
Let be the image in this limit of the coordinate class . These elements are distinct: in , equality would require , whose inner-coordinate section is a singleton and hence not in the nonprincipal . For , the truth of any expanded-language formula on is exactly
This expression depends on the formula and the tuple length, not on the chosen increasing indices. Thus is an order-indiscernible sequence. The empty-coordinate stage embeds elementarily into .
Their Skolem hull is elementary by the witness-function argument, and an order automorphism of extends to it by transporting the generators in every term. Equality and relation preservation follow from the displayed uniform truth test, and the inverse order automorphism gives the inverse map. Therefore this proves the Ehrenfeucht-Mostowski theorem through ultrapowers and their directed limit, with no Ramsey-theoretic step.
The exponent-two Erdős-Rado theorem for pairs says that for every infinite cardinal ,
In particular, a finite or countable coloring of pairs from has an uncountable monochromatic subset. We prove the general statement by an end-homogeneous tree construction.
Let be any coloring on an ordinal . Route a vertex through an increasing sequence of preceding vertices as follows. Start with the least vertex. After ancestors , , have been selected, choose the least remaining candidate , above those ancestors, satisfying
The terminal candidate always qualifies, so the procedure eventually reaches it. Candidate sets shrink and their least elements increase. If an intermediate candidate is , its route is precisely the earlier part of the route to , because all earlier color tests agree. Consequently the routes define a set-theoretic tree on the vertices.
Along a branch, the color from an ancestor to every later branch vertex is constant. Also a node of height is determined by its sequence of colors to its ancestors: reconstruct the ancestors as successive least candidates with those prescribed colors, then reconstruct the node as the least final candidate. Thus its level has at most nodes.
Apply this with and . If every node had height less than , each level would have size at most , and there would be at most such levels. Their total size would be at most , contradicting . Hence some route has at least ancestors. For each ancestor mark its constant color to all later route points. Since there are only colors, one occurs on ancestors. Those ancestors form the required monochromatic set. This proves the theorem.
To make the ordinal function precise, one useful ordinal partition-bound function is
The target here is ordinal order type. For infinite , the proved theorem with supplies a bound : its homogeneous set has order type at least , and it handles all the required color counts. Finite targets are handled by finite Ramsey theorem bounds. Thus is total, nondecreasing and .
Suppose a sequence of transfinite iterates , with suprema at limit stages, has supremum an uncountable cardinal at a limit stage, with every earlier iterate below . Then
choose an earlier iterate at least and use the next one as a bound. If the iteration has already stabilized, it has already found a fixed point instead.
This closure forces to be a strong limit cardinal. For any infinite , color pairs of distinct binary strings of length by their first differing coordinate. This uses colors and has no monochromatic triangle: three binary digits at one coordinate cannot be pairwise different. It gives a coloring on with no homogeneous target of type . Since colors are allowed when the target is the ordinal , we have
Finite exponents satisfy the strong-limit requirement automatically for uncountable .
The needed additional condition is the tree property: every tree of height , with nonempty levels of size less than , has a cofinal branch. In the present convention it forces to be a regular cardinal. If were singular, take a cofinal sequence of smaller cardinals of length and form the disjoint union of chains of those lengths. This tree has height and total size , levels of size at most , but no branch of length , a contradiction.
For any , apply the preceding color-routing tree to a coloring . If a route already has length , use it. Otherwise its levels are indexed below and each has size
for infinite exponents, by the strong-limit property; finite cases satisfy the same needed bound directly. The tree has nodes, so regularity forces its height to be . The tree property supplies a cofinal branch. The ancestor-color partition of that branch has one class of size , since and is regular. Thus for every , and
We have proved that an uncountable function-closed cardinal with the tree property is regular and strong limit, hence strongly inaccessible, and that it is the requested fixed point.
The paper's final sentence, read as a claim about the tree property alone, is false. Already has it by the König infinity lemma but is not uncountable; even with uncountability stipulated, there are relative-consistency models with the tree property at . The latter counterexample is documented in the research account of Mitchell forcing. The rigorous conclusion in the iterated-function context is the one just proved: function closure supplies strong limit, and the tree property supplies regularity and a branch. Strong limit and regularity follow from distinct hypotheses in this context.

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact