WorksMathematics
Root separation in one-parameter families of integer polynomials
Abstract
For every we construct irreducible integer polynomials of degree and arbitrarily large height with two roots at distance . All the roots can be chosen positive, or the two close roots can be chosen nonreal with all the other roots positive. We also construct monic irreducible examples. For a squarefree family of positive degree in , let denote the exponent for which the least distance between two roots of is as . Among families of degree in and degree three in , the maximum of is . It is for families irreducible over , and for monic families of odd degree . For arbitrary degree in , we prove . Two roots in different irreducible factors over have separation exponent strictly less than . If denotes the supremum of over squarefree families of degree in , then tends to .
Use of artificial intelligence. The research was carried out under human direction from the choice of topic and research questions to the selection of results for the paper. OpenAI’s ChatGPT 6.2 Astra developed the constructions and proofs and drafted the manuscript. Anthropic’s Claude Fable 5.2 reviewed the manuscript, checked examples, and suggested improvements to the proofs and to the comparison with known results. The human review did not include an independent check of every proof.
Introduction
Let be squarefree of degree , with complex roots . Its height is the largest absolute value of its coefficients, and its root separation is
For define by . Let be the upper limit of as over squarefree integer polynomials of degree . Write for the corresponding upper limit over irreducible polynomials and for the upper limit over monic irreducible polynomials.
Mahler’s inequality gives [11]. For , Bugeaud and Dujella constructed reducible polynomials giving [6 Theorem 1]. The two close roots lie in different factors, one of which is linear. For irreducible polynomials they obtained for [5 Theorem 1]. Our construction gives
For this improves their bound for irreducible polynomials. It falls short of their reducible bound by .
The construction in § 2 starts with the cubic . For small positive it has two inverse branches near . We construct a rational function so that the two values of on these branches agree to high order. At a fixed large real value , the corresponding two values of are at distance of order , while the height of the polynomial is of order . One sign of gives positive roots. The other gives positive roots and a pair of nonreal conjugates.
Using in § 3, we obtain
These bounds improve Bugeaud and Dujella’s bound for odd and even [6 Theorem 3]. For , Bugeaud and Mignotte proved [7 Theorem 4]. For reducible monic polynomials, Bugeaud and Dujella obtained in even degrees and in odd degrees [6 Theorem 2]. Our irreducible construction falls short of the former by and exceeds the latter by one. In degrees 5, 7, 9, the reducible monic examples of Dujella and Pejković give , , , respectively [8 (3.2)]. Our monic irreducible examples give , 4, in those degrees.
We now consider , squarefree in , of degree in and degree in . Define its separation exponent by
where runs through the positive integers. For all sufficiently large , the specialization is squarefree of degree and has height . Here we use the height of the specialization itself. If the coefficients of in are coprime in , Bézout’s identity bounds their common divisor at every integer by a fixed integer, so passing to the primitive part gives the same exponent. To see that the limit exists, put . By the Newton–Puiseux theorem [15 §3.1], the roots have convergent Laurent expansions in near zero, with coefficients in a number field, for some positive integer . The difference of any two distinct roots has a first nonzero term . Their distance at is asymptotic to , so their separation exponent is . The number is the largest such order .
Let be the supremum of over these families, and write and for the suprema over reducible and irreducible families in , respectively. We call absolutely irreducible when it is irreducible in . If is irreducible over , Hilbert’s irreducibility theorem, applied to its primitive part in , gives arbitrarily large positive integers for which is irreducible over [18 Lemma 4]. Replacing by gives negative integers of arbitrarily large absolute value.
In the notation of Pejković [14], . His results give and [14 Theorem 2.1 and Corollary 3.6]. In degree four he constructs an irreducible family with exponent , and proves for every squarefree family [14 Proposition 3.1 and Theorem 3.5]. In particular, his irreducible quartics improve on the value given by Theorem 2.1 in that degree.
When , Theorem 4.3 gives the exact maxima for all squarefree families and for irreducible families. For odd , Proposition 4.4 gives the monic maximum , attained by an absolutely irreducible family. The main estimate comes from an auxiliary polynomial whose first coefficients vanish when evaluated at one root. The order of the difference between two roots controls how many of the same coefficients vanish at the second root. The degree of the resultant limits the total order of vanishing.
For arbitrary , Zhuang’s bound over gives for [20 Corollary 8.1.2]. Take the place at infinity and replace by . This multiplies both the order of the root difference and the degree in by . Taking th roots and letting removes the multiplicative constant. Theorem 5.1 gives
If the two roots lie in different irreducible factors over , Theorem 5.3 gives for . In the Bugeaud–Dujella examples and equality holds. The two proofs use different properties of the curve . The first bounds how long two branches can agree in terms of the two degrees. For roots of different factors we use the number of intersections of the two curves, together with a theorem of Ruberman.
Finally, Theorem 6.2 shows that and have the same limit, and that this limit equals . Thus a single family of degree with would improve the coefficient in every sufficiently large degree. The same improvement would hold for absolutely irreducible families in infinitely many degrees.
Irreducible polynomials of every degree
Theorem 2.1. For every there is an absolutely irreducible polynomial of degree three in and degree in with the following property. For all real of sufficiently large absolute value, two roots of are at distance . The pair is real for one sign of and consists of nonreal conjugates for the other. The polynomial is irreducible over for infinitely many integers of each sign. For integers of sufficiently large absolute value, its primitive part has height . In particular,
Put . For small positive , the equation has two roots near and a third root near . We construct a rational function , with , for which . Thus vanishes to higher order than as . The same identity for will be used in the monic case.
Lemma 2.2. Let , and , where . For an indeterminate let
The space has a basis consisting of polynomials of degrees and . Any two polynomials , forming a basis are coprime, and for some we have
whenever are the three roots of .
Proof. Put . Setting in the partial fraction expansion of gives
For the degree of is at most . For , we show that the coefficients of and determine the remaining coefficients. Equating the coefficients of in the definition of gives a linear system whose matrix has entries for .
For , expansion of the generating function in powers of gives . For the entry vanishes unless and have the same parity. For and , with , it is
Grouping even and odd indices separately in this case, we obtain in both cases square matrices of the form
The entry counts paths from to with steps and . By the Gessel–Viennot lemma [9], the determinant is a signed count of families of disjoint paths. Only the identity permutation of the endpoints can occur. Indeed, of two disjoint paths, the one starting farther left must be above the other wherever both are defined. If it ended at the lower endpoint, which lies farther to the right, this would already fail when the other path reached the higher endpoint. At least one such family exists. The th path takes diagonal steps followed by horizontal steps. Thus the determinant is positive. The two highest coefficients can now be prescribed arbitrarily, and every nonzero polynomial in has degree at least .
Choose , of degrees , , respectively. For the three roots , , put
This is a symmetric polynomial in , of total degree at most . As and , it is a polynomial in of degree at most . To prove (1), we first compare at two roots. In the equation defining , take first and then . Multiply the first equation by and the second by and subtract. The term indexed by disappears. The two remaining terms give
Here . Cancelling the factors and gives .
Let and choose another root of tending to 1. The equality shows that is bounded at zero. As , this gives with . To see that , take large positive . The three roots of , divided by , tend to the three cube roots of unity. Since for some , its three values are then distinct.
A common zero of , would make divisible by
If this monic quadratic divided , it would equal . Its linear coefficient would then give , for which the constant term is . Thus , are coprime. Changing the basis multiplies the left side of (1) by a nonzero rational constant.
Proof of Theorem 2.1. Apply the lemma with and . Since the basis polynomials are coprime, we may choose a basis , with and . Put . Near choose the square root of with value 1 there. The function has derivative 1 at . By the inverse function theorem we can write , and . Write . Let be the analytic solution of with . As , we have . By (1),
The right side has a zero of order at . Since , the first nonzero term of cannot have degree one or three. A zero of order at least four would give , which is impossible for . For the identity would give with , contrary to the lemma. Hence with .
We now fix and compare the two corresponding values of . Put , taking the square root with value at . Then and . Denote by the value of corresponding to . Thus and . For the difference quotient at two nearby points is . Applying this with and gives
The left side has a zero of order , so has a zero of order . The difference therefore has order in . Its absolute value is of order because . For small real with , both values of are real. For , the two values of are purely imaginary and conjugate. Since the inverse functions have real Taylor coefficients, the two distinct values of are then nonreal conjugates.
Clear the fixed denominators of . Form the polynomial
The coefficients are symmetric polynomials with integer coefficients in the three roots of the monic cubic , so they belong to . The coefficient of is the nonzero product . To see the degree in , factor for any for which the degree is . Since , we obtain
This holds for all but at most one , since . In particular, . Putting in the preceding calculation now gives the asserted distance between two roots of .
Suppose that in . Substituting and gives zero, so one factor, say , vanishes identically after this substitution. For all but finitely many , the three roots of are distinct, and (1) gives three distinct values of . All three are roots of . It follows that . Thus depends only on . For each fixed , the product in (2) is a nonzero polynomial in , since and have no common zero. Hence has no complex zero and is constant. This proves absolute irreducibility.
In particular the coefficients of as a polynomial in are coprime in . By the Euclidean algorithm, some -linear combination of them is a nonzero integer . For integer the greatest common divisor of the coefficients of divides . Since , its primitive part has the same order of height.
Hilbert’s irreducibility theorem, applied to and , gives infinitely many irreducible specializations of each sign. Dividing by gives the lower bound for .
Corollary 2.3. For all sufficiently large , the polynomial in Theorem 2.1 has positive real roots for one sign of , and positive real roots and two nonreal conjugate roots for the other.
Proof. We locate the other roots by showing that has simple zeros in . For , the other two solutions of are and , where
As increases from to infinity, traces an arc in the upper half-plane starting at , with and . The polynomial has no zero on the open arc, since would also give , contradicting (1). The defining equation for gives
We count zeros of the real part on the right by estimating the change in the argument of along this arc.
For a real zero of , the argument of lies between and and tends to . Thus its change is at least , including the case . For a nonreal conjugate pair , the product starts on the positive real axis and has imaginary part . Since decreases strictly, this imaginary part can change sign only once, from positive to negative. The product never vanishes. Its continuous argument starts at zero and tends to or . Summing over the zeros of with their multiplicities, we find that the change in the argument of is at least .
Finally . The second factor has constant argument , and the argument of the first changes from to . Hence the net change of the argument of is at least
By continuity, the argument takes at least distinct values congruent to modulo in the interior of the arc. The real part vanishes at each of these points. It follows that has at least distinct zeros in . Its double zero at accounts for the rest of its degree, so all zeros are simple. For each such zero , the function has a simple zero at , since . Its local inverse gives a real solution of tending to as . The corresponding roots of tend to the distinct positive numbers , as is strictly increasing on . When the remaining two roots are real, they too are positive, since with real and nonzero. □
An irreducible sextic for every odd parameter
For we have . When we may take
For the three roots of , substitution of and gives
Since at zero, the proof of Theorem 2.1 gives a sextic of height with two roots at distance . For large positive this pair is positive. The quartic has alternating signs at , beginning with a positive value. Its four distinct positive zeros give the four remaining roots as in Corollary 2.3. For odd the polynomial reduces modulo two to . Put and . At each root of we have , by direct substitution. Taking the product over the three roots and using their elementary symmetric functions gives
Thus reduces to . It has no root in . The only remaining possible irreducible factors of degree at most three are , and . Indeed, a cubic over is irreducible precisely when it has no root there. Division by the three displayed polynomials gives the respective nonzero remainders , and . Thus the sextic is irreducible modulo two. It follows that is irreducible over for every odd integer .
Monic irreducible polynomials
Theorem 3.1. For every there is an absolutely irreducible , monic of degree in and of degree three in , with the following property. For all sufficiently large positive , all roots of are real and the distance between two of them is , where
There is also an absolutely irreducible family, again monic in and of the same degrees, for which a pair of nonreal conjugate roots is at distance for all sufficiently large positive . In both families the specialized polynomial is irreducible over for infinitely many positive integers . Consequently .
Proof. It suffices to arrange that . Clear the denominators of and . The leading coefficient of in is then a nonzero integer independent of . By (2), the leading coefficient of in is . Thus is monic in and has integer coefficients.
Take and let be odd. By Lemma 2.2, has a basis consisting of an odd polynomial of degree and an even polynomial of degree . Indeed, the even and odd parts of each element of belong to , and every nonzero element has degree at least . At the defining equation for gives . These zeros are simple. Otherwise , which is impossible when . For , the identity would give with , contrary to the lemma. Since and are coprime and , we have and .
Let be the three solutions of near , numbered so that , and . Put . By (1),
Since has a simple zero, vanishes to order at zero. Both and have simple zeros there. Their derivatives at zero are therefore equal and nonzero. This common derivative is rational, so we may multiply by a rational constant to make it one. If and , then . As , the quotient tends to . Since has a simple zero in , vanishes to order in . Taking and gives the asserted distance for odd .
For even take and replace by
These polynomials have degrees and and are coprime. Indeed, . At , the equalities and give and hence . At every other zero of we have . On the two inverse branches the values of are . Both have derivative at zero, and their difference is
This difference vanishes to order . Since both derivatives are nonzero, the inverse functions give the asserted distance as in the odd case.
We next show that all roots are real for sufficiently large . In both constructions is odd, so for a polynomial of degree . For the other two solutions of are , where . Set
As increases, describes an arc in the upper half-plane from to infinity. Its real part decreases strictly and its argument tends to . By (1), has no zero on the open arc. The estimate in the proof of Corollary 2.3 therefore gives . The defining equation for gives
The argument of changes from to . Hence the change in the argument of the quotient is at least . Its real part vanishes at least times. Thus has at least distinct zeros in . Their negatives and account for all zeros of , so each is simple. In even degree has the additional simple zero . Since the denominator has degree one more than the numerator, (or ) also has a simple zero at infinity, as can be seen by putting and expanding at . For sufficiently small real , each of these simple real zeros gives one real solution of (or ). Their images under are the roots of the polynomial in (2).
Write in odd degree and in even degree. In both cases is odd. Hence and have rational coefficients and the same degrees as and . Indeed, every odd rational function can be written as for some rational function . The two zeros of at give inverse branches interchanged by complex conjugation. For real near zero, their -values are distinct complex conjugates. The substitutions multiply both coordinates by nonzero constants, so the difference vanishes to the same order as before.
For odd , identity (1) shows that takes three distinct values at the three roots of for all but finitely many . The same holds for even , since the map is one-to-one for wherever it is defined. The substitutions using preserve this property. Thus the factorization argument following (2) proves absolute irreducibility in all cases. Hilbert’s irreducibility theorem gives infinitely many positive integer specializations that are irreducible over . The change of scale at the beginning of the proof multiplies all root distances by and preserves the separation exponent. □
Sharp bounds for families of degree three in the parameter
Let have degree in and degree in . Divide the coefficients of , considered as a polynomial in , by their greatest common divisor in , and denote the quotient by . Put . If , all the roots are constant and . Otherwise, choose an integer at which the coefficient of in does not vanish. With and , the polynomial
has degree in , since is nonzero by primitivity. The coefficient of in is the nonzero value used to choose , and is nonzero because . Thus has degree in , and neither nor divides it. The substitutions are invertible when . Since neither nor divides , the irreducible factors of correspond to those of , over as well as over . All roots of stay bounded as , since has degree . For two roots we have
If , take a pair for which the difference in has order . The transformed difference has the same order for a finite common limit, and greater order at a common pole. In either case its order satisfies .
For later use, let be squarefree of degree in and in , with . Its discriminant with respect to is a nonzero polynomial of degree at most in , by homogeneity in the coefficients of . All roots of are bounded at zero, where the leading coefficient of is nonzero. Hence
We recall the two properties of the resultant used below. If has leading coefficient and roots , then
For nonzero , the resultant vanishes identically exactly when have a common factor of positive degree in . Otherwise its degree in is at most . Indeed, each term of the Sylvester determinant uses coefficients of and of .
Lemma 4.1. Let be irreducible, of degree in and in , and suppose that has degree . Let be the order at of the difference of two roots. If , then is an integer and
Proof. Since , (3) shows that only one pair has order . The roots have convergent expansions in for a common . Multiplication of by permutes the roots and preserves the order of every difference. Hence it fixes or interchanges the two chosen roots.
For an integer consider the polynomials of degree at most in and at most in . Their coefficients form a vector space of dimension . For every nonzero such , irreducibility of and give
If both roots are fixed, only integral powers of occur, so is an integer. Choose so large that . We can take a nonzero so that the first coefficients of vanish. Since is divisible by , the order of is at least . All other factors in the product for are bounded at zero. Thus , giving in this case.
::: If the roots are interchanged, write them as and , where . Indeed, two applications of the substitution fix each root, so only powers of occur. The first odd term of has degree . In every series the coefficients of vanish. Take so large that . Requiring the series to vanish to order at least therefore imposes at most homogeneous linear conditions. Choose a nonzero solution . The product then vanishes to order at least in . Hence , which gives .
For every , Lemma 4.1 gives the exact maximum for the separation exponent of an absolutely irreducible cubic family of degree in the parameter. Indeed, the order is at most . To attain it, take from Theorem 2.1 with and put
The polynomial is absolutely irreducible and divisible by neither variable. Interchanging the variables and taking reciprocals therefore gives an absolutely irreducible of degrees in , respectively. In the notation of the construction, , and the two roots of have difference of order in . Hence .
Lemma 4.2. Let be squarefree over , of degree in and degree in . Then
Proof. We may assume and use the bounded coordinate . For two roots of one irreducible factor, Lemma 4.1 bounds the order of their difference by . For roots of two distinct factors of degrees and , the resultant gives . If one factor is independent of , it is linear in and . Otherwise all four degrees are positive and
Since and , both cases give the asserted bound.
Theorem 4.3. Let be squarefree over , with and . Then
If is irreducible over , then
Both upper bounds are attained for every .
Proof. The first bound is Lemma 4.2 with . For the second assume that is irreducible over and . Choose two roots in the bounded coordinate whose difference has order . If they lie in the same irreducible factor over of degrees , Lemma 4.1 gives
If they lie in distinct factors, recall that the primitive polynomial in is irreducible over . Its irreducible factors over are conjugate over . To see this, substitute , where . The monomials in each factor give distinct powers of , so the factor can be recovered from the resulting univariate polynomial. Normalize the latter to be monic. Its coefficients are algebraic numbers, being symmetric functions of roots of a polynomial in . Thus each normalized bivariate factor has algebraic coefficients. The product of the distinct Galois conjugates of such a factor is rational and hence exhausts the factorization of . If there are factors, each has degrees . Since divides , we have . The resultant then gives .
The general bound is attained by the Bugeaud–Dujella families for [6 proof of Theorem 1]. Their construction also covers . Put and . For the family of degree is , where
In both cases . The positive root of tends to zero, and uniformly between and . Thus , as required. Theorem 2.1 gives equality in the irreducible case for all . □
Proposition 4.4. Let be squarefree over , monic of degree in , and of degree three in . Then
For every odd , equality is attained by an absolutely irreducible family.
Proof. Put . Suppose that two roots have difference of order greater than in . If they lie in one irreducible factor over of degrees , Lemma 4.1 gives the bound after the change to . It exceeds only if . For roots in different factors the resultant bound is at most three if one factor is independent of . Otherwise their degrees can be ordered so that and . Then , which can exceed only for and . Both factors can be taken monic in . The second then has a polynomial root of degree two in , whereas every root of the first is by Cauchy’s bound. It remains to consider an absolutely irreducible .
Let be the order after the change to . Lemma 4.1 and its proof show that and that the two expansions are , with . The first odd term of has degree . To see that is a local parameter on this branch, let be the Taylor polynomial of through degree . Then has order one in . We have . Indeed, if had a zero of order in , the original difference would have order in .
We show that and are rational functions of a single parameter. In the function field , a nonconstant function has zeros and equally many poles, counted with multiplicity on all branches, including those at infinity [19 Theorem 1.4.11]. Thus and have and three poles, respectively, counted with multiplicity. Every nonzero polynomial of degrees at most has at most zeros, since its total pole order is at most . The space of these polynomials has dimension , and . Take with . Requiring to vanish to order at least imposes at most conditions, since its coefficients of already vanish. Choose two independent solutions . Their ratio is nonconstant because . For all but finitely many , the zeros of lie away from the chosen branch and the poles of . They are additional zeros of . Hence
which proves .
Every pole of lies over , since is monic in . If there were two distinct poles, both would give . Choose a nonzero linear combination of vanishing at . It would have at least zeros at the chosen branch and one at each pole, since on that branch. This exceeds . Take the unique pole of as . Since , is a cubic polynomial in . At the chosen branch has first odd term of degree , and has order at most three. Its order is therefore two. Affine changes of and now give , with the chosen branch at . Write in lowest terms. This rational function has degree , a double zero at , and a zero at infinity. In particular .
Take the three solutions of with and as . The two values of near have difference of order in . Indeed, in the parameter we have with . At fixed the two parameter values satisfy . Subtracting the expressions for gives , as required. As in Lemma 2.2,
is a polynomial of degree at most . Since has order in , and , we obtain with . Multiply the numerator identity by and sum cyclically over . The left side cancels. Using on the right gives
Thus belongs to the space of Lemma 2.2 with complex coefficients and vanishes at . Its basis polynomials are coprime, so this condition defines a one-dimensional subspace. The proof of Corollary 2.3 gives a polynomial of degree spanning that subspace. This contradicts . Theorem 3.1 gives equality for odd . □
Upper bounds for arbitrary degrees
Theorem 5.1. Let be squarefree over , of degree in and degree in . Then . For we have
A curve with local equation has a singularity of type for . Let be the largest that occurs on a plane curve of total degree . Gusein-Zade and Nekhoroshev proved [10]. Orevkov raised the lower bound to [13 §5]. Astashov also constructed examples with prescribed degrees in each variable [1 Theorem 4]. We need an upper bound in terms of those two degrees that also applies when other branches pass through the same point.
Lemma 5.2. Let be squarefree, with degrees in , respectively. Suppose that in analytic coordinates centred at a point of , its equation has a factor , where . Then
Proof. Make a fractional linear change in each variable, keeping the given point finite. Choose the new fibres at and in each variable to avoid the singular points, to meet the curve transversely and to have all four of their crossings off the curve. This is possible since a reduced curve has only finitely many exceptional fibres in each projection. After clearing denominators, the two degrees are still . All four corner coefficients of the rectangle are nonzero, and the coefficients of and of have only simple nonzero
roots in the other variable. The polynomial is therefore convenient and nondegenerate at infinity. It follows that it is cohomologically tame [17 p. 199], hence spectrally tame [4 Proposition 3.2.5]. The zero fibre has only isolated singularities because it is reduced. We can thus apply the spectral inequality of [4 Theorem 4.3.1].
For and put . Form a list containing and for each pair, together with copies of . For any irrational , the spectral inequality gives
where entries are counted with multiplicity. On the right, is the spectrum at infinity in the convention with entries in , by [12 Theorem 5.16]. The list can also be read directly from [3 Table 1, $e=0$] by counting the pairs with each given value of .
On the left, are the entries for [2 p. 252]. If the local equation is with , replace by . For small the equation at the origin has just the indicated factor. Local semicontinuity [4 Theorem 4.1.1] therefore bounds its contribution by that of the original equation.
Take irrational, with smaller than and than every nonzero . All local entries lie in . The global list has entries. By its symmetry about , we must subtract one entry for each and one for each . The total number subtracted is
This gives the asserted bound. □
Proof of Theorem 5.1. Suppose . Use the polynomial of Section 4, with and . Its degrees are and , and its roots are bounded at . Let be the largest order of their pairwise differences. Then . For , the discriminant bound (3) gives and hence .
Let . If , then . Otherwise (3) shows that exactly one unordered pair attains . Continuation around preserves this pair. Its mean and its squared difference are therefore convergent power series in . Write the latter as with an integer and . Taking a local square root of , the coordinates and give the factor . Lemma 5.2 gives in this case as well. Since and , the result follows. □
For roots in different factors, the resultant bounds the order of their difference by the total number of intersections of the two curves, counted with multiplicity. When the order exceeds , a small translation splits the contact into simple intersections. A second bound follows by enclosing these intersections in a single disc and applying a theorem of Ruberman.
Theorem 5.3. Let be squarefree over , of degree in and in . If two roots belong to different irreducible factors over , the order of their difference in is strictly less than . For it is at most .
Proof. We may assume that the order is positive. Pass to the bounded coordinate of Section 4, and denote the order of the transformed difference by . It is at least the original order. Let be the degrees of the two factors in , and let be their degrees in , respectively. Put
Their resultant has degree at most in . Its order at zero is the sum of the orders of all differences between a root of and a root of , since both leading coefficients are nonzero there. In particular, and . If , then and . A factor independent of is linear in and again . These cases give both asserted bounds whenever applicable. We may therefore take and . Then , so suffices.
Suppose that . Since , the chosen pair is the only one from the two factors whose difference has order at least . Continuation around zero preserves each factor and hence fixes each root of the pair. Write the two roots as convergent power series . In particular, is an integer. Also , so we may label the factors so that and .
We first separate the chosen pair from all the other roots on a disc containing intersections. Write with . Choosing a holomorphic -th root of near zero gives a local coordinate in which . Translate the curve by in the -coordinate. The graphs of and now meet at the solutions of . Put . In the annulus , these two graphs are at distance at most in each fibre. Every other root of either factor is farther away. Indeed, before translation its difference from has order . For a root of this follows from uniqueness of the pair. For another root of apply uniqueness to its difference from and use . Its distance from on the annulus is therefore , uniformly on its Puiseux branches. As , this is arbitrarily large compared with , even after the translation.
Choose sufficiently small. Cut the annulus along a ray between two successive solutions and take a closed disc with smooth boundary in the cut annulus containing all solutions in its interior. Over the tube contains just the two chosen graphs, one from each curve. We can now perturb the coefficients of both equations slightly and generically, keeping their bidegrees. In the resulting curves are smooth, connected and transverse to each other by Bertini’s theorem for the two systems of positive bidegree. Inside the tube each still has one simple root over every point of . Their difference stays nonzero on and, by Rouché’s theorem, still has zeros in . There are intersections in all, by the intersection formula for curves of the two bidegrees.
Reverse the orientation of the second curve. Remove the two graph discs over and join their boundaries by the straight segment between the two roots in each fibre over . These segments form an embedded annulus. The tube is free of other sheets, and the two boundary orientations agree with this gluing. The surface is now connected. At each of the remaining transverse intersections take local complex coordinates in which the original sheets are the axes and the second is the -axis. After its orientation is reversed, the local smoothing , for small , joins the two sheets by an oriented annulus. Replacing two small discs by this annulus adds one handle to the connected surface. A general curve of bidegree projects to the -line with degree and simple branch points, counted by its homogeneous discriminant. Riemann–Hurwitz gives its genus , and likewise for the second curve. Thus the resulting closed embedded oriented surface has genus
The first replacement takes place in the contractible tube over , and each later one in a small ball. Thus the projection degrees of the resulting surface are the differences and . By Reuberman’s theorem [16 Corollary 1.3], an embedded surface with nonzero projection degrees has genus at least . Consequently
This gives . For example, when the two bounds for a contact of order are and . Adding to yields , as required. □
Proposition 5.4. Let be as in Theorem 5.1. Suppose two roots belonging to distinct irreducible factors over have difference of order greater than in . Both factors can then be chosen over , and both roots have convergent Laurent expansions in with rational coefficients. Consequently a pair in distinct factors with at least one root nonreal for arbitrarily large positive has difference of order at most .
If is irreducible over and has irreducible factors over , then .
Proof. Use the bounded coordinate of § 4 and let be the parameter degree after removing the common factor of the coefficients. The transformed pair has order . The discriminant bound (3) gives a total order at most , so every other pair has smaller order. Continuation around zero fixes the unordered pair and preserves each factor. The two roots therefore have ordinary power series in .
The coefficients of these series are algebraic numbers by the Newton–Puiseux theorem [15 §3.1]. Every conjugation over preserves the order of each difference and thus fixes the unordered pair. If a conjugation interchanged its roots, the two factors would have the same degrees in . Their resultant would give , since and . Thus each root and its factor are fixed. The coefficients of the series and of the factors, after one coefficient of each factor is scaled to one, are rational. The formula gives the asserted Laurent expansions, which take real values for all sufficiently large positive .
For the last assertion the factors are conjugate over , as in the proof of Theorem 4.3, so each has degrees . For roots in different factors the resultant bounds the order by . For roots of one factor (3) bounds it by . Division by gives the result. □
From one degree to all large degrees
Orevkov used power substitutions and a linear change of variables to produce -singularities on plane curves of increasing total degree [13 Proposition 5.2]. Here we increase the degree in while keeping the degree in fixed. The order of a single root difference gives a lower bound in the new degrees.
Proposition 6.1. Let be squarefree over , with and . For every sufficiently large there is a reducible squarefree family of degree in and degree in with exponent at least
If is irreducible over , the same estimate holds for absolutely irreducible families of degree in along an infinite arithmetic progression of degrees . If , then has an absolutely irreducible factor with .
Proof. Put and use the bounded coordinate of § 4. The resulting polynomial has degrees , where , and its leading coefficient in is nonzero at . Its coefficients in have no common zero in , since the original equation was divided by the common factor of its coefficients. Choose a pair of roots of whose difference has order , and denote the corresponding roots of by . Then
By (3), the sum of the orders of all pairs is at most . Every other pair therefore has order strictly less than . Applying any conjugation over to the algebraic coefficients of the Puiseux expansions preserves all orders and thus fixes the unordered pair . In particular its common value at is rational. Continuation around also fixes the pair, so it either fixes or interchanges its two members. Thus and have power series in .
If belong to the same absolutely irreducible factor, every conjugation fixes that factor once a coefficient is normalized to one. Its coefficients are algebraic, as in the proof of Theorem 4.3, and hence rational. When , Theorem 5.3 shows that the two original roots do belong to the same factor. Its primitive equation divides , and . The two roots give .
Choose such that has degree and has only simple roots. There are only finitely many excluded values. Put
For any even integer with , set
After choosing a square root of near zero, the two functions have convergent power series . Let be their local inverses. Then , and
The last expression has order in . Taking gives two roots of at distance .
To compute the degrees, put . The coefficient of in is , of degree . The highest power of in is therefore , with nonzero constant coefficient. In particular . Also , by the substitution for . It follows that has degrees in and in , and that .
The polynomial is squarefree over because the discriminant of is nonzero at the transcendental element . Its coefficients have no common zero in . For this follows from the corresponding property of . At the polynomial specializes to times the nonzero coefficient of in . Thus it is squarefree in . The substitution is invertible, and replacing by preserves squarefreeness after denominators are cleared since . After multiplication by a nonzero integer, belongs to and has exponent at least .
Given a sufficiently large , take the largest even for which . Then satisfies . Multiplication by distinct factors , with integers chosen outside the constant roots of , produces a reducible squarefree family of the required degrees. Its exponent is at least .
Finally suppose is irreducible over . Since , Proposition 5.4 shows that is absolutely irreducible. Take a root of and put . At a simple root of the implicit function theorem gives an embedding . The polynomial is Eisenstein at , so it is irreducible over and hence over . If , it follows that . As , we obtain . Thus is irreducible over . Its coefficients have no common zero, so it is irreducible in . The same changes of variables as above now give the absolute irreducibility of . Taking proves the estimate along the stated progression. □
Theorem 6.2. We have
The common value belongs to . It exceeds if and only if some absolutely irreducible family satisfies . Moreover, as if and only if for every .
Proof. Write . Since and Theorem 5.1 gives , we have . If , choose a family of some degree with . Proposition 6.1 gives . Since , the first two limits equal .
The absolutely irreducible examples of Theorem 2.1 give . If , choose and a family of degree with . Proposition 6.1 gives an absolutely irreducible factor with . Applying the same proposition to gives absolutely irreducible families in infinitely many degrees with . Letting tend to proves (4). Conversely, every absolutely irreducible family with gives by definition.
If , the first limit in (4) gives . The definition of then gives for all , and the converse follows at once. □
Funding. T. P. acknowledges support from the Croatian Science Foundation under project IP-2022-10-5008 (TEBAG). He also acknowledges support from the project “Implementation of cutting-edge research and its application as part of the Scientific Center of Excellence for Quantum and Complex Systems, and Representations of Lie Algebras” (Grant No. PK.1.1.10.0004), co-financed by the European Union through the European Regional Development Fund under the Competitiveness and Cohesion Programme 2021–2027. This research was funded by the European Union through NextGenerationEU under the National Recovery and Resilience Plan 2021–2026 (institutional grant of the Faculty of Science, University of Zagreb, IK IA 1.1.3. Impact4Math).
References
References
- [1]E. A. Astashov, A_k-singularities of curves and surfaces of fixed degree, quasi-degree, or multi-degree, Moscow Univ. Math. Bull. 70 (2015), 241–247.
- [2]M. Borodzik, M. Hedden and C. Livingston, Plane algebraic curves of arbitrary genus via Heegaard Floer homology, Comment. Math. Helv. 92 (2017), 215–256.
- [3]M. Borodzik and T. K. Moe, Topological obstructions for rational cuspidal curves in Hirzebruch surfaces, Michigan Math. J. 65 (2016), 761–797.
- [4]M. Borodzik, A. Nemethi and A. Ranicki, On the semicontinuity of the mod 2 spectrum of hypersurface singularities, J. Algebraic Geom. 24 (2015), 379–398.
- [5]Y. Bugeaud and A. Dujella, Root separation for irreducible integer polynomials, Bull. Lond. Math. Soc. 43 (2011), 1239–1244.
- [6]Y. Bugeaud and A. Dujella, Root separation for reducible integer polynomials, Acta Arith. 162 (2014), 393–403.
- [7]Y. Bugeaud and M. Mignotte, Polynomial root separation, Int. J. Number Theory 6 (2010), 587–602.DOI
- [8]A. Dujella and T. Pejković, Root separation for reducible monic polynomials of odd degree, Rad Hrv. Akad. Znan. Umjet. Mat. Znan. 21 (2017), 21–27.
- [9]I. Gessel and G. Viennot, Binomial determinants, paths, and hook length formulae, Adv. Math. 58 (1985), 300–321.
- [10]S. M. Gusein-Zade and N. N. Nekhoroshev, Singularities of type A_k on plane curves of a chosen degree, Funct. Anal. Appl. 34 (2000), 214–215.
- [11]K. Mahler, An inequality for the discriminant of a polynomial, Michigan Math. J. 11 (1964), 257–262.DOI
- [12]Y. Matsui and K. Takeuchi, Monodromy at infinity of polynomial maps and Newton polyhedra (with an appendix by C. Sabbah), Int. Math. Res. Not. IMRN (2013), no. 8, 1691–1746.
- [13]S. Yu. Orevkov, Some examples of real algebraic and real pseudoholomorphic curves, in Perspectives in Analysis, Geometry, and Topology, Progress in Mathematics, vol. 296, Birkhäuser, 2012, 355–387.DOI
- [14]T. Pejkovic, Real and p-adic root separation for cubic and quartic polynomials, preprint, https://web.mathematik.uni-kiel.de/~pejkovic/parsep/rootsep34-v1.pdf.
- [15]A. Poteaux and M. Rybowicz, Good reduction of Puiseux series and applications, J. Symbolic Comput. 47 (2012), 32–63.DOI
- [16]D. Ruberman, The minimal genus of an embedded surface of non-negative square in a rational surface, Turkish J. Math. 20 (1996), 129–133.
- [17]C. Sabbah, Hypergeometric periods for a tame polynomial, Port. Math. (N.S.) 63 (2006), 173–226.
- [18]A. Schinzel and U. Zannier, The least admissible value of the parameter in Hilbert’s irreducibility theorem, Acta Arith. 69 (1995), 293–302.DOI
- [19]H. Stichtenoth, Algebraic Function Fields and Codes, 2nd ed., Graduate Texts in Mathematics, vol. 254, Springer, Berlin, 2009.DOI
- [20]W. Zhang, Symmetric Diophantine approximation over function fields, Ph.D. thesis, Leiden University, 2015.