Define the normalized progression average
We first prove the needed four-term progression bound with torsion factors, valid for the arbitrary cyclic modulus in the question:
For the factor at slope zero, write , , and apply Cauchy-Schwarz to discard :
For concise bookkeeping put , the conjugate of the preceding derivative convention, and . Expanding the square with and setting gives . Another Cauchy-Schwarz, now in , discards and yields
Expand with , set , and apply Cauchy-Schwarz a third time to discard . The remaining -average is a full average over , so
Negating does not change the average. The nonnegative integrand depends only on and . Averaging over the subgroups and costs at most their indices and compared with unrestricted increments. The unrestricted expression is exactly , obtained by expanding the last squared average. Thus the bound is .
Reversing the progression handles slope three. To isolate slope one, eliminate slopes three, zero and two successively in the same calculation; the surviving increments have coefficients , so only the index is needed. Reversal handles slope two. This proves the minimum-over-factors bound, without assuming that or is invertible modulo .
If , the requested lower bound is zero and is immediate. Otherwise let . Telescoping the product of four copies of against the constant product gives four error terms, each with one factor and three factors bounded by one. Hence
In particular, choosing gives
These are ordered modular progression parameters , as requested; small-order differences and are included. A claim of distinct nontrivial progressions would need a separate subtraction and size condition.
Use and the unnormalized Fourier transform
We state the eighth-power convention explicitly. A bounded function is -quadratically uniform if
where means conjugation and every expectation is uniform. The expression is real and nonnegative. Applying Parseval identity on a finite group to the multiplicative derivatives gives the equivalent condition
This is the Derivative identity for the Gowers U3 norm. A set of density is quadratically uniform when its balanced subset indicator satisfies that condition, not when its unbalanced indicator does. Some definitions bound rather than its eighth power; that norm parameter is in the convention here. Quadratic uniformity controls three-dimensional additive cubes and four-term progression counts; small ordinary linear Fourier coefficients alone are not the same condition.
Let count ordered solutions in , including the trivial ones. The hypothesis gives . Its relation to additive energy is the promised easy Fourier step. Translate into and embed it in an odd prime cyclic group of order . No relevant integer equality wraps modulo . With the unnormalized Fourier transform , character orthogonality gives
Since multiplication by permutes the characters, Cauchy-Schwarz and Parseval identity on a finite group imply
Thus . Equivalently, one can apply Cauchy-Schwarz directly to .
By part (i), some has and . Part 3(ii) then bounds . Apply part (ii) with : there is , , Freiman 2-isomorphic to a subset with . Consequently has a positive subset density depending only on , and tends to infinity with because .
View the representatives of as a subset of . The assumed four-term case of Szemerédi's theorem gives four distinct integer representatives in arithmetic progression once is sufficiently large in terms of . If their preimages are , the two model equalities
lift through the Freiman 2-isomorphism to the same equalities in . Hence their consecutive differences agree; injectivity makes the common difference nonzero. Therefore contains a nonconstant four-term arithmetic progression. This proves dense three-term progressions force four-term progressions without an assumption that initially lies in a short interval.
For , the following precise cardinality-controlled cyclic Freiman model suffices. If is nonempty, and has , there is with
Here means a bijection preserving and reflecting all equalities of -term sums, a Freiman s-isomorphism of order . The same construction works for if only injectivity is needed.
Choose an auxiliary prime larger than and every nonzero absolute value in , and large enough to embed the finite set injectively modulo . Multiply its residues by a uniformly chosen nonzero residue . For a fixed , is uniform among the nonzero residues. The forbidden residues are
of number at most . The union bound says that the probability some nonzero enters is at most . Fix a multiplier avoiding every forbidden event.
Let represent . Partition that range into half-open intervals of length and retain the most populated interval; it supplies with . Define the model map by . For two -term sums of retained representatives, their difference has . If the original sums are equal, then , hence , so the model preserves the equality.
Conversely, if the model sums are equal, then is a multiple of . If their original difference were nonzero, it would satisfy with , contradicting the forbidden-event choice. If directly, implies too by the choice of . Thus the equality is reflected. Padding equal one-term images with copies of a fixed retained element shows that the map is injective. This proves the Freiman s-isomorphism and the claimed bound on the cyclic modulus, independently of the diameter of .
One quantitative form of the Balog-Szemerédi-Gowers theorem is: if a nonempty finite set in an abelian group has and additive energy , where , then there is with and both and bounded by . The constants depend only on . We prove explicit, nonoptimal polynomial bounds, thereby proving the required Balog–Szemerédi statement.
Let . Then . Declare to be the set of popular sums. The remaining sums contribute at most . Since , it follows that . Also . Form a bipartite graph on two copies of , joining to when . Its density is at least .
We prove the needed four-step path lemma for a dense bipartite graph. Call an ordered pair of left vertices bad if its number of common right neighbors is less than , where . For a uniformly chosen right vertex , let and let count bad ordered pairs in . Then
Consequently some satisfies
For this , and . Delete vertices having more than bad partners in . At most half are deleted; the retained set has . For any , at least middle vertices are good partners of both. Each such supplies at least walks of four edges. Hence every pair in has at least
such walks. Repetition of vertices is allowed; it does not affect the counting argument.
Each walk has edge labels
all in , and . For a fixed pair , these four labels determine , so distinct walks have distinct label quadruples. For each element of , fix one representing pair . Label quadruples for distinct differences cannot overlap. Therefore
Using and yields
This proves the small-difference-set form of the Balog-Szemerédi-Gowers theorem. To obtain a small sumset too, write and apply the already proved Plünnecke inequality to starting set and summand set . Some nonempty has . For any fixed , , whence . Thus the conventional small-sumset conclusion also follows with constants depending only on .
Apply Plünnecke's inequality with the summand set also equal to . There is a nonempty satisfying for every .
We need the Ruzsa triangle inequality in a form whose short proof also fixes the signs. For every , choose one representation . The map
is injective into : its difference recovers , then the fixed representation recovers . Therefore
Use and and the common set from the previous part:
Thus
The shared minimizing subset is important; applying separate growth estimates with unrelated subsets would not justify this bound.
The sumset form of Plünnecke's inequality states: for nonempty finite sets in an abelian group, if , there is a nonempty such that
Here . The same can be used for every . This sumset statement follows from an elementary minimal-growth argument.
Choose minimizing among nonempty subsets of . Then and for every , including the empty set. We prove the Petridis minimal-growth lemma
for each finite set . Enumerate . Put and define
The disjoint new contributions to have sizes , so . For , the point was already present in , hence . Consequently the new contribution at stage to has size at most
Summing proves the lemma. Starting with and then taking gives inductively , proving the stated Plünnecke inequality.
Let be the ordinary Hilbert class field of . Its real places split completely, so is totally real. Since is quadratic and is totally imaginary, cannot equal ; hence it equals . Therefore the compositum has
and is an abelian extension of , by restriction of the Galois group to . At finite primes it is unramified because unramified extensions remain unramified under base change. The field has no real places, so there is no infinite-place ramification condition left. Thus , where is the Hilbert class field of . Since , the tower degree formula proves class number divisibility for a CM extension:
This argument uses the ordinary class field with split real places; replacing it silently by the narrow class field would invalidate the claim that is totally real.
The hypotheses make a CM-field, with nontrivial automorphism over the totally real number field . In every complex embedding of , this automorphism acts as complex conjugation. Write , and for the finite cyclic group of roots of unity in .
For any , the element is an algebraic integer, indeed a unit, and every conjugate has absolute value one. The Kronecker theorem on algebraic integers in the unit disk makes it a root of unity. For clarity, the essential proof is finite: the conjugates of all positive powers have modulus one, so the integer coefficients of their monic minimal polynomials are bounded by fixed binomial coefficients. There are only finitely many such polynomials and roots; two powers coincide, forcing a root of unity.
Thus
is a homomorphism. Its kernel is exactly , since a conjugation-fixed element lies in and both it and its inverse are integral there. For , and . Consequently induces an injection
To check injectivity explicitly, if , then , so . The cyclic group has even order because it contains , and its quotient by squares has order two. Therefore the unit index of a CM-field satisfies
For number fields, a prime ideal divides the different ideal exactly when it is ramified. More precisely, the local different exponent and tame ramification relation is , with equality for tame ramification; if , the finite residue extension is separable and the exponent is zero. One can see the ramification criterion directly from the trace pairing modulo the base prime: when its quotient ring has a nonzero nilpotent ideal, whose elements pair to zero since multiplication by a nilpotent has trace zero. When the quotient is a product of finite separable fields and its trace pairing is nondegenerate. Its determinant is a unit exactly in the second case.
Let . Any rational prime ramifying in must ramify in each , because ramification indices multiply in a tower. It would then divide both field discriminants, contrary to their coprimality. Thus is unramified at every finite prime, so . The Minkowski lower bound for a number-field discriminant excludes this for a field of degree :
The last inequality follows already for the weakest signature bound : the expression starts above one at and increases with . Consequently .
For completeness, normality converts this intersection statement into the desired degree statement. Put . Restriction injects into . The fixed field of its image in is exactly , because is Galois. The Galois correspondence therefore gives . This proves coprime discriminants imply linear disjointness:
Use , , , and denote their relative codifferents by , and . Transitivity of the field trace gives
For the forward direction, testing an element against every is legitimate because . For the reverse direction, take and use trace transitivity.
Localize at a nonzero base prime ideal. The intermediate ring is semilocal Dedekind domain, so its invertible fractional ideal is principal there, say locally. This principality follows by choosing, with the Chinese remainder theorem for ideals, a generator at each of its finitely many maximal ideals. The displayed criterion now becomes
Invert these fractional ideals using Solution 1. Equality at every base localization gives the global different in a tower formula
The second factor is extended to the upper ring of integers; that extension is implicit in the printed formula. At a local prime the exponents consequently satisfy .
Set and , with valuation rings . Let be the ramification index, the residue-field degree, and the residue characteristic. A base uniformizer generates in , up to a unit. The free -module has rank . For an integral , its field trace is the matrix trace of multiplication by on this module. Reduction modulo preserves the matrix trace.
The quotient has a filtration
Each of the successive quotients is a one-dimensional vector space over . On each, multiplication by acts as multiplication by its residue , whose trace over is . A basis adapted to the filtration makes the multiplication matrix block triangular, so its trace is the sum of these quotient traces. This proves trace reduction in a ramified local extension:
Since the hypothesis says , the right side is zero in . Thus
No normality hypothesis is needed. The original PDF only requires divisibility of , not divisibility exactly once; the converted TeX has corrupted that condition and several local subscripts.
Use , the valuation ring of the completion of a number field at a prime ideal, with uniformizer and residue field . This completeness is essential for coprime-factor Hensel lifting. Normalize the two residue factors to be monic first; if the original factors have reciprocal leading units, undo the normalization by multiplying the eventual factors by reciprocal lifts of those units.
Let the two monic residue degrees be . Start with monic lifts satisfying . Suppose monic lifts of these degrees satisfy . Put , of degree less than . Over , the linear map
is injective: a zero image makes divide , by coprimality, forcing by its degree bound, and then . The dimensions agree, so this map is a bijection. Choose lifts whose image is , and set
Their product agrees with modulo ; the new cross term has factor and . The degree bounds preserve monicity. Each coefficient is a Cauchy sequence in the complete discrete valuation ring , so the limits satisfy
If were interpreted only as an ordinary localization rather than the completed valuation ring, the assertion would be false. For example splits modulo seven but is irreducible over . The completion notation in the next part provides the intended local-field interpretation.
Let . Each field automorphism preserves the ring of integers and the base prime ideal, so it permutes the prime ideals above . Suppose there were more than one orbit, and fix one orbit . By the Chinese remainder theorem for ideals, choose such that
The field norm lies in . It lies in every prime ideal of , since for a fixed member each conjugate of is zero modulo that member. Thus its contraction lies in . On the other hand, for outside , every is also outside it, so every equals one modulo . The product is consequently one modulo . This contradicts its membership in . Therefore the Galois action on the primes above is transitive. The normal extension of number fields is separable, so its conjugates are exactly the elements of used in the product.
For pairwise comaximal ideals of , reduction gives the Chinese remainder theorem for ideals:
Its kernel is . For two comaximal ideals, write with , . If , then , showing . Induction gives the product formula: a product of ideals individually comaximal with is still comaximal with , by multiplying identities modulo .
Put . Since , choose with . The element has any prescribed residues , proving surjectivity. Its residue class modulo the product is unique because we already identified the kernel. In a Dedekind domain, powers of distinct nonzero prime ideals are pairwise comaximal, giving the prime-power form. The theorem requires comaximality; it is not a statement about arbitrary lists of ideals.
For an auxiliary field varying slowly over lattice distances, measure in lattice-spacing units and replace the squared difference by . The long-wavelength action becomes
This is the Euclidean path integral of a one-coordinate quantum particle, with the classical spatial coordinate serving as imaginary time and . To see the correspondence at the operator level, time-slice it by and use the symmetric kernel
Acting on a smooth test function, Gaussian integration gives . Thus the continuum auxiliary-field transfer operator for an Ising chain is , where
Multiplying kernels and integrating over intermediate fields gives for fixed endpoints, and for periodic fields, up to the normalization already absorbed in . This is the precise classical Ising chain as a quantum particle correspondence. It concerns an imaginary-time transition kernel; it is not the squared modulus of a real-time transition amplitude. The continuum approximation retains smooth long-distance modes, rather than asserting exact equality of the finite-spacing transfer operator with this differential operator.
At , set . Expansion of the local potential gives
The potential is even and grows as at large . It is a symmetric double well precisely when
The nonzero stationary points satisfy . For , the ratio decreases strictly from 2 to 0: writing , the sign follows from , whose derivative is . Hence there is one minimum on each side when . At the potential instead has a single quartic minimum, and for a single quadratic minimum. Small with fixed positive falls in the double-well regime, but the printed claim requires this coupling condition; alone does not imply it if scales to zero too quickly.
A weak magnetic field gives
The linear term is odd in . For positive , the positive well becomes deeper. If the zero-field minima are , their depth difference to first order is . The magnetic field tilts the symmetric double well and favors the matching spin orientation. A large enough field can eliminate the metastable well.
Finally, two minima do not imply a true finite-temperature phase transition in this one-dimensional chain at fixed finite interaction range . A finite quantum barrier permits quantum tunnelling; the confining one-dimensional Schrödinger operator has a unique symmetric ground state at zero field and a nonzero first excitation gap. The quantum gap corresponds to a finite classical correlation length. A very small gap can mimic ordering on a finite sample, but genuine degeneracy requires a singular limit such as vanishing tunnelling. This distinction is central to interpreting the quantum mapping.
Use the infinite translation-invariant chain first. Set and Fourier-transform the exchange matrix:
For positive this is strictly positive. Its reciprocal is a linear polynomial in , proving the nearest-neighbour form of the exponentially coupled Ising inverse kernel:
Expanding the squared difference gives
Since , substitution into part (a) yields
This is the exact bulk identity. A finite periodic version uses the periodized coupling ; its discrete Fourier symbols are the same and the squared differences wrap around the ring. Simply truncating the exponential at open ends gives a different inverse at those ends.
For an open chain of sites with the literal finite matrix , the open-chain endpoint corrections for an exponential Ising kernel are explicit:
The inverse has endpoint diagonal entries , interior diagonal entries and adjacent entries . Multiplication by the original geometric matrix verifies every entry. These boundary terms affect boundary amplitudes, but not the bulk free-energy density in the thermodynamic limit at fixed .
Write the finite-chain spin vector as and the exchange matrix as . For and this matrix is positive definite, as can be seen from the positive Fourier symbol in part (b), or from the explicit positive open-chain inverse. The linear-source multivariate Gaussian integral, obtained by completing the square, gives
Indeed the exponent is . Insert this Hubbard–Stratonovich transformation into the spin partition function. The field term remains , and each independent spin sum is
Consequently
The interaction convention counts the full ordered pair sum, with no factor of one-half, so the auxiliary coupling is . Diagonal terms only contribute a constant to the spin Hamiltonian; retaining them makes the given positive exchange matrix and normalization convenient. The auxiliary field is not itself the physical spin magnetization: the conditional spin expectation is , so in the auxiliary-field integral.
The renormalization group organizes how a statistical system changes when observed at progressively larger length scales. Start with its partition function and a microscopic cutoff . Split the Fourier transform field into modes below and modes in the shell . Integrate the shell modes, rescale coordinates to restore the cutoff, and rescale the field to keep a chosen gradient normalization. This is a renormalization-group transformation. It preserves the partition function when the generated interactions and additive free energy terms are retained; truncating the resulting operator expansion is an approximation.
A renormalization-group fixed point has an invariant dimensionless interaction structure. The continuous flow of a nearby scaling field takes the linearized form
A positive defines a relevant operator, a negative an irrelevant operator, and a zero a marginal operator requiring nonlinear analysis. The continuous-flow eigenvalues should not be confused with the discrete multipliers . To reach a critical point, tune the relevant thermal variable and set the relevant symmetry-breaking source to zero. The resulting flow stays on the critical surface.
Microscopically different systems whose remaining couplings approach the same fixed point share a universality class. Irrelevant perturbations lose their influence on long-distance exponents. Relevant directions explain why varying the temperature or field moves the system away from criticality. The thermal scaling field obeys , while the rescaled correlation length is . Taking therefore gives .
If the field has scaling dimension , the source term has eigenvalue . The singular free-energy density then obeys
when the hyperscaling relation is valid. This derives the scaling hypothesis for critical phenomena and the exponent relations of Question 2 from the fixed-point picture.
At the Gaussian fixed point, keeping invariant gives . The quadratic coupling has and the quartic coupling has . Thus quartic interactions are relevant below four dimensions, marginal at four, and irrelevant above four. Below four, in dimensions where an ordinary continuous short-range scalar transition exists, an interacting Wilson-Fisher fixed point governs that transition. Above four, the Gaussian fixed point gives mean-field powers, but the quartic term remains necessary to stabilize the ordered state: it is a dangerously irrelevant coupling, explaining the failure of naive hyperscaling. The renormalization group connects scale invariance, universality, relevant tuning parameters and fluctuation corrections in one framework.
The Ginzburg criterion tests whether fluctuations on the scale of one correlation volume are small compared with the squared equilibrium order parameter. In the scalar quartic Landau-Ginzburg theory of part (a), the ordered mean-field value is , and its longitudinal correlation length is proportional to .
Average the fluctuation field over a region of size . Its smoothing factor selects momenta of order or smaller. The Gaussian correlation function then gives, up to a positive dimension-dependent constant,
The last step follows directly by setting . Dividing by proves the correlation-volume derivation of the scalar Ginzburg ratio:
For , the ratio diverges as the transition is approached, so there is a fluctuation-dominated critical region of a phase transition in which mean-field critical exponents cannot be trusted. With , an estimate of its width is
in fixed microscopic units. Its numerical coefficient depends on the averaging convention and microscopic normalization.
For , the ratio tends to zero and the long-distance mean-field approximation becomes self-consistent. Four is therefore the ordinary quartic upper critical dimension. At this test is marginal: its thermal power is zero, and the running quartic interaction produces logarithmic corrections. The marginal Ginzburg criterion alone cannot decide those logarithms. Short-distance contributions to an unaveraged variance renormalize the coefficients; they should not be mistaken for the long-wavelength fluctuations used in this criterion. Nor does the criterion by itself prove that a proposed ordered phase exists below its lower critical dimension.

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