Complete intersections on general hypersurfaces
E.Carlini, L.Chiantini, A.V.Geramita

TL;DR
This paper investigates the conditions under which complete intersections of a given codimension can be contained in a general hypersurface in projective space, providing a complete characterization for certain codimension ranges.
Contribution
The authors give a complete answer to when complete intersections of codimension r can lie on a generic hypersurface in projective space for the case 2r ≤ n+2, based on degrees.
Findings
Complete characterization for when complete intersections lie on general hypersurfaces for 2r ≤ n+2.
Conditions depend on degrees of hypersurfaces and generators.
Provides a comprehensive answer to a classical geometric problem.
Abstract
We ask when certain complete intersections of codimension can lie on a generic hypersurface in . We give a complete answer to this question when in terms of the degrees of the hypersurfaces and of the degrees of the generators of the complete intersection.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsCommutative Algebra and Its Applications · Polynomial and algebraic computation · Algebraic Geometry and Number Theory
Complete intersections on general hypersurfaces
Enrico Carlini
Dipartimento di Matematica, Politecnico di Torino, Torino, Italia
,
Luca Chiantini
Dipartimento di Scienze Matematiche e Informatiche, Università di Siena, Siena, Italia
and
Anthony V. Geramita
Department of Mathematics and Statistics, Queen’s University, Kingston, Ontario, Canada, K7L 3N6 and Dipartimento di Matematica, Università di Genova, Genova, Italia
[email protected] ](mailto:[email protected]%20)
Abstract.
We ask when certain complete intersections of codimension can lie on a generic hypersurface in . We give a complete answer to this question when in terms of the degrees of the hypersurfaces and of the degrees of the generators of the complete intersection.
1991 Mathematics Subject Classification:
Primary ; Secondary
1. Introduction
Many problems in classical projective geometry ask about the nature of special subvarieties of some given family of varieties, e.g. how many isolated singular points can a surface of degree in have? when is it true that the members of a certain family of varieties contain a line? contain a linear space of any positive dimension? The reader can easily supply other examples of such questions.
This is the kind of problem we consider in this paper: what types of complete intersection varieties of codimension in can one find on the generic hypersurface of degree ?
In case it was known to Severi [Sev06] that for the only complete intersections on a general hypersurface are obtained by intersecting that hypersurface with another.
This observation was extended to by Noether (and Lefschetz) [Lef21, GH85] for general hypersurfaces of degree . These ideas were further generalized by Grothendieck [Gro05].
Our approach to the problem mentioned above uses a mix of projective geometry and commutative algebra and is much more elementary and accesible than, for example, the approach of Grothendieck. We are able to give a complete answer to the question we raised for complete intersections of codimension in which lie on a general hypersurface of degree whenever .
The paper is organized in the following way: in the next section (Section 2) we lay out the question we want to consider and explain what are the interesting parameters for a response.
In Section 3 we collect some technical information we will need about varieties of reducible forms and their joins. In order to find the dimensions of these joins (using “Terracini’s Lemma”) we calculate the tangent space at a point of any variety of reducible forms. We also recall some information about artinian complete intersection quotients of a polynomial ring.
In Section 4, we use the technical facts collected in Section 3 to reformulate our original question. We illustrate the utility of this reformulation to discuss complete intersections of codimension in on a general hypersurface when . We further use our approach to give a new proof for the existence of a line on the general hypersextic of .
In Section 5 we state and prove our main theorem which gives a complete description of all complete intersections of codimension in which lie on a generic hypersurface when .
2. Question
The objects of study of this paper are complete intersection subschemes of projective space. Recall that is a complete intersection scheme if its ideal is generated by a regular sequence, more precisely, , , and form a regular sequence in . If for all , we will say that such a is a and we will assume ; notice that is unmixed of codimension in . With this notation we can rephrase the statement
the degree hypersurface contains a
in terms of ideals as follows
for forms forming a regular sequence and such that for all .
Clearly not all choices of the degrees are of interest for us, e.g. if for all , then no can be found on a degree hypersurface. On the other hand, any hypersurface of degree contains a if , for some . Simply cut that hypersurface with general hypersurfaces of degrees .
So, one need only consider where none of the .
Lemma 2.1**.**
Let with . The following are equivalent facts:
- •
there is a on the general hypersurface of degree in ;
- •
there is a on the general hypersurface of degree in .
Proof.
Let , where is a general hypersurface of degree in . Let be the ideal of a , with degrees as above.
Then if and only if
[TABLE]
and hence, if and only if , where is the complete intersection defined by . ∎
From this Lemma it is clear that the basic question to be considered is:
(Q): For which degrees does the generic degree hypersurface of contain a ?
If, rather than restricting to the generic case, we asked if some hypersurface of degree contains a , then the answer is trivial. Indeed, the ideal of any () always contains degree elements.
3. Technical facts
If is a partition of the integer (i.e. and ) we write . For each we define a subvariety (where ) as follows:
[TABLE]
We call the variety of reducible forms of type . The dimension of is easily seen to be . (For other elementary properties of see [Mam54] and for the special case see [Car06], [Car05] or [Chi02] for the case).
If are independent points of we will call the spanned by these points the join of the points and write
[TABLE]
More generally, if are varieties in then the join of is
[TABLE]
In case we write
[TABLE]
and call this the * (higher) secant variety of *.
Joins and secants of projective varieties are important auxiliary varieties which can help us better understand the geometry of the original varieties (see e.g. [CGG02, CGG03, CC06, Cil01, Ger96, LM04, Ådl88, CJ96]). One of the most fundamental questions we can ask about joins and secants is: What are their dimensions?
This is, in general, an extremely difficult question to answer. The famous Lemma of Terracini (which we recall below) is an important observation which will aid us in answering this question.
Lemma 3.1** (Lemma of Terracini).**
Let be reduced subvarieties of and let be a generic point of .
Suppose that , then the (projectivized) tangent space to at , i.e. , can be described as follows:
[TABLE]
Consequently,
[TABLE]
We want to apply this Lemma in the case that the are all of the form . A crucial first step in such an application is, therefore, a calculation of where .
Proposition 3.2**.**
Let , and let be a generic point of .
Write where and let be the ideal defined by:
[TABLE]
Then the tangent space to at the point is the projectivization of and hence has dimension
[TABLE]
Proof.
Consider the map of affine spaces
[TABLE]
defined by
[TABLE]
Let be the point . A tangent direction at is given by any vector of the form and the line through in that direction is
[TABLE]
A simple calculation shows that the tangent vector to at the point is exactly and that proves the proposition.
∎
In view of Terracini’s Lemma, the following corollary is immediate.
Corollary 3.3**.**
Let all be partitions of where
[TABLE]
Let
[TABLE]
be an ideal of generated by generic forms where
[TABLE]
If
[TABLE]
then
[TABLE]
Remark 3.4*.*
It is useful to note the following facts:
- (i)
In Proposition 3.2 we are using the fact that has characteristic 0. The problem is that the differential is not necessarily generically injective in characteristic . 2. (ii)
Observe that the generic point in , can always be written as the product of irreducible forms with the property that any -subset of these forms () is a regular sequence. 3. (iii)
This last can be extended easily to joins of varieties of reducible forms. I.e. the generic point in such a join can be written as a sum of elements with the property that each summand is a point enjoying the property described in (ii) above. Moreover, every -subset () of the set of all the irreducible factors of all of these summands is also a regular sequence. 4. (iv)
Fröberg (see [Frö85]) has a conjecture about the multiplicative structure of rings , where and is an ideal generated by a set of generic forms. This conjecture gives the Hilbert functions of such rings. However, apart from the cases (proved several times by various authors, see [Frö85, GS98, IK99]) and (proved by [Ani86]) this conjecture has resisted attempts to prove it.
Notice that in terms of the geometric problem in Corollary 3.3, one need only consider Fröberg’s conjecture for a strongly restricted collection of degrees.
We will need some specific information about the Hilbert function of some artinian complete intersections in polynomial rings. The following lemma summarizes the facts we shall use.
Lemma 3.5**.**
Let and be generic forms in having degrees
[TABLE]
and
[TABLE]
for a non-negative integer .
Consider the quotient
[TABLE]
and its Hilbert function . The following facts hold:
- (i)
* is symmetric with respect to ;* 2. (ii)
if then is non-increasing for ; 3. (iii)
the multiplication map on given by (the class of in ) has maximal rank.
If one of the following holds
[TABLE]
[TABLE]
[TABLE]
[TABLE]
we also have that:
- (iv)
if , then . 2. (v)
if , then . 3. (vi)
if , then if and only if .
Proof.
As is a Gorenstein graded ring (i) follows immediately, while (iii) is a consequence of a theorem of Stanley [Sta80] and Watanabe [Wat87].
To prove (ii) we can use the Weak Lefschetz property, i.e. multiplication by a general linear form has maximal rank, e.g. see [MMR03]. The condition on , coupled with the Weak Lefschetz property, yields that every element of is the product of a fixed linear form with a form of degree . Now consider an element of , call it , then since is a standard graded algebra, , where is the class of in , and is the class of a form of degree . By what we have seen, where is the form we had earlier and the are forms of degree . Rewriting we get . But is in hence and hence the dimension cannot increase. Proceeding by induction we prove the statement.
As for (iv), it is enough to give the proof for as there are no generators of degree smaller than . Let be a quotient obtained when all the forms and have the same degree . Notice that it is enough to show the result for . In fact, whenever we pass from to another quotient by increasing the degrees of forms we obtain
[TABLE]
and the inequality is preserved; these Hilbert function estimates use the fact that the forms and do not have linear syzygies. By straightforward computations one gets
[TABLE]
and
[TABLE]
Thus the inequality is equivalent to
[TABLE]
Notice that if (2) holds for the pair then it holds for all the pairs with . By direct computations we verify that the inequality is satisfied for and for and . Hence, (2) holds for
[TABLE]
[TABLE]
[TABLE]
To complete the proof of (iv) it is enough to evaluate (1) for in the case , i.e. .
To show (v), notice that by (ii), if
[TABLE]
then is definitely non-increasing and hence it cannot be symmetric with respect to by (iv).
To get (vi) it is enough to use symmetry and (v).
∎
4. Equivalences
In this section we give some equivalent formulations of our basic question (Q), formulated at the end of Section 2.
Clearly, if is a hypersurface of degree and is a , then the ideal inclusion yields
[TABLE]
for forms of degrees . But the converse is not true in general. If and the forms do not form a regular sequence, then is not the ideal of a complete intersection. To produce an equivalence we need to use joins:
Lemma 4.1**.**
The following are equivalent:
- (i)
a generic hypersurface of degree of contains a , where for all ; 2. (ii)
the join of the varieties of reducible forms fills the space of degree forms in variables, i.e.
[TABLE]
Proof.
The implication (i)(ii) simply follows from the ideal inclusion argument above yielding the presentation for the generic degree form, where for all . The implication (ii)(i) is easily shown using the description of the generic element of the join, see Remark 3.4 (iii). ∎
Remark 4.2*.*
Notice that there is an equality of varieties
[TABLE]
for all non negative integers and . Hence, by Lemma 4.1, the condition
[TABLE]
is equivalent to the statement
a generic hypersurface of degree of contains a , where or for all .
It follows from these observations that we can further restrict the range of the degrees in our basic question (Q), i.e. it is enough to consider
[TABLE]
Now we exploit Terracini’s Lemma and the tangent space description given in Corollary 3.3 in order to produce another equivalent formulation of question (Q).
Lemma 4.3**.**
The following are equivalent:
- (i)
the generic hypersurface of degree of contains a , where for all ; 2. (ii)
let and be generic forms in of degrees and respectively, then
[TABLE]
where denotes the Hilbert function in degree of the ring.
Proof.
The condition on the join in Lemma 4.1 can be read in term of tangent spaces as equivalent to
[TABLE]
for generic points . Using the description of the tangent space to the variety of reducible forms this is equivalent to saying
[TABLE]
where is the degree piece of the polynomial ring and the forms and are generic of degrees and respectively. ∎
As a straightforward application we get the following result:
Proposition 4.4**.**
The generic degree hypersurface of contains no , for all , when .
Proof.
We use Lemma 4.3. Consider in generic forms and of degrees and respectively. If we let be the ideal , then we want to show that and for that it is enough to show that is not an artinian ring. As has height and the quotient cannot be zero dimensional and the conclusion follows. ∎
Remark 4.5*.*
Using Lemma 4.3 we can also recover many classical results in an elegant and simple way. More precisely, we can easily study the existence of complete intersections curves, e.g. lines and conics, on hypersurfaces.
Example 4.6**.**
As an example we prove the following without using Schubert calculus:
The generic hypersextic of contains a line.
Proof.
Let and consider the ideal
[TABLE]
where the forms are linear forms and the forms have degree . We want to show that . Clearly
[TABLE]
It is well known [GS98, IK99, Frö85] that 4 general binary forms of degree 5 generate and we are done. ∎
For more on this topic see Remark 5.5.
5. The Theorem
We are now ready to prove the main theorem of this paper, a description of all the possible complete intersections of codimension that can be found on a general hypersurface of degree in when .
Theorem 5.1**.**
Let be a generic degree hypersurface, with . Then contains a , with , and the all less than , in the following (and only in the following) instances:
- •
: then , arbitrary and and can assume any value less than ;
- •
, : for we have that and can assume any value less than ;
- •
, : for we have that and can assume any value less than ;
- •
* or : for we have that can assume any value less than ;*
- •
* or , or : we have only linear spaces on quadrics, i.e. and .*
Proof.
Recall that from Lemma 2.1 and Remark 4.2 it is sufficient to consider the existence of a on the generic hypersurface of degree when .
When , by Proposition 4.4, we know that no complete intersection exists. Hence we have only to consider the cases and
In order to use Lemma 4.3 we consider the generic forms and of degrees and respectively. If we let and we want to check whether or not.
The case . In this case, is an artinian Gorenstein ring and is the first place where one has . Thus, the generic degree hypersurface contains a if and only if and this is equivalent to the inequality
[TABLE]
which is never satisfied unless and .
The case will be proved using Lemma 3.5. In order to do this, we divide the proof into three parts:
- •
the hyperplane case: any ;
- •
the plane case: for , and hence ;
- •
the four space case: and for , and hence , ;
- •
the general case: one of the following holds
[TABLE]
The hyperplane case. We need to study on the generic degree hypersurface of . As one of the generators of the complete intersection is a hyperplane, we can reduce to a smaller dimensional case. In algebraic terms, for a generic linear form , we consider the surjective quotient map
[TABLE]
to get the following:
the generic element of can be decomposed as a product of forms of degrees , i.e. it has the form with and ,
if and only if
the generic element of can be decomposed as a product of forms of degrees , i.e. it has the form with , .
Hence, we have to study on the generic degree hypersurface of , i.e. codimension complete intersections in , . As this situation was treated before and the only case where the complete intersections exist is for and for all .
The plane case. We have to study on the generic degree curve of for , any such that . Now, and we consider forms and of degrees, respectively, and . We want to study the ring
[TABLE]
and to compare and in order to apply (iii) of Lemma 3.5 to show that
[TABLE]
Using Lemma 3.5 (i), we see that the last non-zero value of occurs for
[TABLE]
In particular, for , and a exists for any and , . If , then and the same conclusion holds for . Finally, if , then and it is easy to see that . Hence, for and any such that , the generic degree plane curve contains a .
The four space case. We address the case of on the generic degree 4 threefold in . Hence, we consider and generic quadratic forms and . Let be the quotient ring
[TABLE]
and notice that by the vanishing of the lefthand side of (2) in the proof of Lemma 3.5 we have that . Applying (ii) of Lemma 3.5 yields and hence the required exists.
The general case. Consider the ring
[TABLE]
and the multiplication map given by the degree form
[TABLE]
Clearly, with this notation, one has that the generic degree hypersurface contains a if and only if
[TABLE]
and this is equivalent to the surjectivity of . We also recall that by Lemma 3.5 (iii) has maximal rank. Hence, to study the surjectivity we only have to compare and .
When , all the degrees are equal to 1, and this was treated in the hyperplane case.
Now we consider the case and we let . If then by Lemma 3.5(v), thus cannot be surjective. Standard computations yield
[TABLE]
while for the inequality never holds. Thus we get that, when one of the conditions (3) holds, is not surjective if
[TABLE]
or
[TABLE]
In the case we have to be more careful and the distances and have to be considered. When one of the conditions (3) holds, by 3.5(vi), is surjective if and only if . Thus we solve the inequality . This is equivalent to
[TABLE]
while for we always have .
Summing up all these facts one gets that when one of the conditions (3) holds:
[TABLE]
[TABLE]
[TABLE]
[TABLE]
and using the treatment of the hyperplane, the plane and the four space cases we obtain the final result.
∎
Remark 5.2*.*
The fact that a general hypersurface of degree in cannot contain a complete intersection of any type with is also a consequence of a result about vector bundles proved by Mohan Kumar, Rao and Ravindra (see [MKRR06]).
In the existence statement for is classical. The cases in and can be obtained using Theorem 12.8 in [Har92] (see also Proposition 5.6 in this paper). In , for and the existence also follows from the analysis of arithmetically Cohen-Macaulay rank two bundles on hypersurfaces, contained in [AC00] and [Mad00].
In , for the case , when , the result also follows from the existence of a canonical curve on the generic quintic threefold of , as was essentially proved in [Kle00].
Remark 5.3*.*
For the theorem above states that the generic degree plane curve contains a for any , but it does not say that this is a set of points. The complete intersection scheme could very well not be reduced. Actually, we can show reducedness and hence the following holds
the generic degree plane curve contains complete intersection points for any .
Remark 5.4*.*
Again a remark in the case . Taking the theorem above states that is the whole space. Now, quite generally, the points of the variety of secant lines either lie on a true secant line or on a tangent line to . We claim that Proposition 3.2 allows us to conclude that the points of the tangent lines are in fact already on the true secant lines. In fact, if , then any point of a tangent line to can be written as for forms of degrees respectively, and scalars and . Thus, lies on the secant line to joining and . In conclusion, we can rephrase the equality in terms of polynomial decompositions
let , any degree form in three variables can be written as for suitable forms of degree and of degree .
This answers a question raised during a correspondence between Zinovy Reichstein and the first author.
Remark 5.5*.*
The restriction in Theorem 5.1 is related to the fact that Fröberg’s conjecture is only known to be true, in general, when the number of forms does not exceed one more than the number of variables. However, there are other partial results on this conjecture that we can use to extend our theorem. E.g. in [HL87] Hochester and Laksov showed that a piece of Fröberg’s conjecture holds. More precisely they showed that if an ideal is generated by generic forms of the same degree then the size of that ideal in degree is exactly what is predicted by Fröberg’s conjecture. Using this we can prove the following:
Proposition 5.6**.**
The generic degree hypersurface in contains a complete intersection of type where or for all i, if and only if
[TABLE]
When this is the well known result on the non-emptyness of the Fano variety of -planes on the generic degree hypersurface of (e.g. see [Har92] Theorem 12.8).
Proof.
Using Lemma 4.3 we have to show the vanishing, in degree , of the Hilbert function of
[TABLE]
for generic linear forms and generic degree forms . Clearly, as the linear forms are generic, we have
[TABLE]
Hence if and only if contains all the degree forms. Using the result by Hochester and Laksov this is equivalent to
[TABLE]
∎
Example 5.7**.**
The variety of reducible quartic hypersurfaces of and its secant line variety provide interesting examples for several reasons.
First note that is a variety of dimension . From Corollary 3.3 it is easy to deduce that . Thus, is a defective variety whose virtual defect is,
[TABLE]
Consider the Noether-Lefschetz locus of quartic hypersurfaces in with Picard group . The quartic hypersurfaces which contain a line are clearly in the Noether-Lefschetz locus. If is a line defined by the linear forms then the form , of degree 4, defines a hypersurface containing if and only if
[TABLE]
I.e. if and only if . Since, as we observed,
[TABLE]
this forces the secant variety to be a component of the Noether-Lefschetz locus.
We wonder how often joins of other varieties of reducible forms give components of the appropriate Noether-Lefschetz locus.
Since is defective for secant lines we have, by a theorem of [CC06], that for every two points on there is a subvariety , containing those two points, whose linear span has dimension , where is the defect of .
We now give a description of such ’s for the variety .
Let be two points of and let be the line in defined by . Consider , the subvariety of reducible quartics whose linear components contain . Clearly . Notice that the linear span of , , is contained in the subvariety of all quartics containing and that variety has dimension .
Thus,
[TABLE]
as we wanted to show.
Notice that the existence of , as above, gives another proof of the defectivity of .
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[AC 00] Enrique Arrondo and Laura Costa. Vector bundles on Fano 3-folds without intermediate cohomology. Comm. Algebra , 28(8):3899–3911, 2000.
- 2[Ådl 88] B. Ådlandsvik. Varieties with an extremal number of degenerate higher secant varieties. J. Reine Angew. Math. , 392:16–26, 1988.
- 3[Ani 86] David J. Anick. Thin algebras of embedding dimension three. J. Algebra , 100(1):235–259, 1986.
- 4[Car 05] Enrico Carlini. Codimension one decompositions and Chow varieties. In Projective varieties with unexpected properties , pages 67–79. Walter de Gruyter Gmb H & Co. KG, Berlin, 2005.
- 5[Car 06] Enrico Carlini. Binary decompositions and varieties of sums of binaries. J. Pure Appl. Algebra , 204(2):380–388, 2006.
- 6[CC 06] Luca Chiantini and Ciro Ciliberto. On the concept of k 𝑘 k -secant order of a variety. J. London Math. Soc. (2) , 73(2):436–454, 2006.
- 7[CGG 02] M. V. Catalisano, A. V. Geramita, and A. Gimigliano. Ranks of tensors, secant varieties of Segre varieties and fat points. Linear Algebra Appl. , 355:263–285, 2002.
- 8[CGG 03] M. V. Catalisano, A. V. Geramita, and A. Gimigliano. Erratum of the publisher to: “Ranks of tensors, secant varieties of Segre varieties and fat points” [Linear Algebra Appl. 355 (2002), 263–285; MR 1930149 (2003 g:14070)]. Linear Algebra Appl. , 367:347–348, 2003.
