REVIEW 2 major objections 4 minor 6 references
A note on $r$ hypersurfaces intersecting in $\mathbb{P}^r$
T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read This paper proves that, for every choice of degrees, the unique largest component of the locus of r-tuples of homogeneous forms with positive-dimensional common zero set—apart from the subfamily where the first r-1 forms already meet in…
desk verdict Genuine extension of Tseng's earlier theorem to all degrees, but the load-bearing codimension bound is outsourced to [Tse18] with an unverified modification. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The proof is a codimension count carried out on incidence correspondences. The central object is the stratum $\Phi^{\mathbb{P}^r,a}_{d_1,\dots,d_k}(\mathbb{P}^r, \operatorname{Span}(r,b))$: tuples whose common zero set contains an integral subscheme of dimension $r-k+a$ whose linear span is exactly a $b$-dimensional plane. Lemma 10, quoted from earlier work, bounds the codimension of this stratum (away from the already-degenerate locus) from below by $-\dim G(b,r)$ plus the minimum, over index sets $i_1<\cdots<i_{r-b+a}=k$, of $\sum_j h_{b,b-i_j+j}(d_{i_j})$, where $h_{r,a}(d)=(r-a)\binom{d+a-1}{d-1}+\binom{d+a}{d}$ is the Hilbert-function lower bound for a nondegenerate $a$-dimensional scheme. For $b=1$ this gives the codimension $-\dim G(1,r)+\sum_i(d_i+1)$ of the common-line family. The numerical heart of the paper is the elementary inequality showing that for every $b>1$ this lower bound strictly exceeds the common-line codimension, so no higher-span stratum can beat the common-line family.
What would settle it
To test Theorem 9, compute directly the codimension of a span-$b$ stratum for a specific degree vector and compare it with the common-line locus. For example, with $r=4$, $a=1$, and $d_1=\cdots=d_4=2$, the theorem predicts the family of four quadrics in $\mathbb{P}^4$ sharing a line has codimension $6$, while the bound for any stratum whose common intersection contains a component with span dimension $b\ge 2$ is at least $7$; exhibiting an explicit $b\ge 2$ family with codimension $6$ or less would refute the claim of uniqueness.
Extended reading notes
Core claim
The central claim is Theorem 9, with Theorem 1 as the case $a=1$: if $a\ge 1$ and $1\le d_1\le \cdots \le d_{r+a-1}$, then the locus $\Phi^{\mathbb{P}^r,a}_{d_1,\dots,d_{r+a-1}}(\mathbb{P}^r)$ of tuples whose common zero set has dimension at least $1$ has a unique component of maximal dimension outside the locus where the first $r+a-2$ forms already have intersection dimension at least $2$, and that component consists of tuples $(F_1,\dots,F_{r+a-1})$ for which $\{F_1=\cdots=F_{r+a-1}=0\}$ contains a line. In the original setting $a=1$, this says that once the degenerate families are set aside, the largest family of $r$ hypersurfaces with positive-dimensional common intersection is always the family of hypersurfaces all containing a fixed line, with no restriction on the degrees $d_i$. The proof works over an algebraically closed field of arbitrary characteristic and, unlike the earlier theorem, requires no inequalities linking the degrees.
Load-bearing premise
The whole argument rests on a codimension bound quoted from earlier work whose key assertion—that at the relevant steps a hypersurface must contain a nondegenerate component of the previous intersection—is only sketched here, and the proof's numerical comparison collapses if that assertion fails.
Editorial extensions
If this is right
- For every degree vector $(d_1,\dots,d_r)$, the common-line family is the unique largest component of the positive-dimensional intersection locus once the subfamilies where the first $r-1$ forms already meet in dimension at least two are set aside.
- The statement extends to $r+a-1$ forms for any $a\ge 1$: adding extra hypersurfaces does not change the dominant mechanism, and the largest component outside the already-degenerate locus is still the common-line family.
- When all degrees are equal, permuting the hypersurfaces recovers the earlier result that tuples vanishing on a line form the unique maximal component.
- The $k=2$, $(d_1,d_2)=(2,2)$ example shows the scope: for $k<r$ the analogous question fails, because the hyperplane family and the quadric family are separate components whose relative sizes depend on $r$.
Reading between the lines
- A natural next step, not taken in the paper, is to identify the second-largest nondegenerate component; the span $b=2$ strata are the obvious candidates, but the gap between the Lemma 10 lower bound and the true codimension is not computed here.
- The same codimension-comparison strategy could be tried on parameter spaces of sections of vector bundles with positive-dimensional common zeros, where a similar trivial family (all sections vanish on a fixed linear space) may dominate; this would transpose the argument beyond hypersurfaces.
- For $k<r$, the paper's $(2,2)$ example shows the dominance of the common-linear-space family is a threshold phenomenon; a testable extension is whether Question 3 holds for all $k<r$ when the common degree is sufficiently large, with the linearly dependent family as the only competitor.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the locus Z of r-tuples of homogeneous forms of degrees d1 ≤ ... ≤ dr in P^r whose common zero locus is positive-dimensional. It claims (Theorem 1) that for all degree choices, the unique maximal-dimensional component of Z not contained in the locus where the first r−1 forms have intersection dimension at least 2 consists of r-tuples of forms all vanishing on some line. The proof generalizes to the loci Φ^{P^r,a} and is carried out in Theorem 9. The key input is a codimension lower bound for the span-b strata, stated as Lemma 10, which is quoted from the author's prior work [Tse18] and only given an informal proof here. The paper also contains a useful example in Section 2 showing that the analogous answer for k < r is not always positive, and a clean numerical comparison in the proof of Theorem 9.
Significance. If the main theorem is correct, it resolves the all-degrees case of a natural question in the geometry of hypersurface intersections, extending previous results of the author and of Slavov. The statement is crisp and the numerical comparison is elegant. The example in Section 2 correctly illustrates that the positive answer fails for proper subsets of hypersurfaces when degrees are equal. The main limitation is the reliance on Lemma 10, whose decisive index condition is not proved in this note; the paper itself flagging this gap is a strength in honesty but also the central weakness in completeness.
major comments (2)
- [§4, Lemma 10 and the Remark following it] The proof of the main theorem rests on Lemma 10, and in particular on the minimum being taken over index sets with i_{r−b+a}=k. This condition is stronger than the condition i_a ≤ k proved in [Tse18, Lemma 4.2], and the manuscript explicitly states in the Remark following Lemma 10 that the existence of indices i_1<...<i_a=k is the part that 'can be made more precise' and is deferred to [Tse18]. The informal argument in the proof of Lemma 10 does not justify why the last instance can be taken to be exactly k; if the backwards induction in [Tse18] only yields i_a ≤ k, then the bound (10) is not established. Because Theorem 9's numerical comparison uses the stronger condition to place the largest degree in the final term, this gap is load-bearing. The author should provide a complete proof of Lemma 10 as stated, or explicitly prove the modified index condition within this paper.
- [§4, proof of Theorem 9] The inequality chain following display (11) relies on the final index being i_{r−b+a}=r+a−1, since it is this term that contributes the summand (b d_{r+a−1}+1) and hence enables the comparison of the lower bound (10) with the line-locus codimension (8). If the index condition in Lemma 10 were weakened to i_a ≤ r+a−1, the last term in the minimum would not necessarily involve the largest degree, and the displayed chain of inequalities would no longer follow. Thus the proof of Theorem 9 is incomplete unless Lemma 10 is fully justified with the stated final-index condition.
minor comments (4)
- [§2, final paragraph] The phrase 'the the answer' contains a duplicated article and should read 'the answer'.
- [Abstract and §1] The notation for the affine spaces of forms, such as \binom{r+d_i}{d_i}, is not typeset consistently; the abstract in particular would benefit from proper display of the binomial coefficients.
- [§4, Lemma 10] The reference to [Tse18] in the proof of Lemma 10 is not specific enough to indicate which statement there is being adapted to the stronger index condition; the author should state precisely which result in [Tse18] implies the required bound, or give a self-contained proof.
- [§2, Table 1] The table of codimensions for the (d1,d2)=(2,2) example is informative, but the text says the computations are verified by an incidence correspondence without giving details; a brief derivation of at least one entry would improve readability.
Circularity Check
No significant circularity: the central all-degrees theorem is not assumed, and the outsourced Lemma 10 is an independent prior result, not a self-fulfilling input.
full rationale
The paper's main claim (Theorem 1, valid for all degree choices) is genuinely new relative to [Tse18, Theorem 1.3], which had degree restrictions. The proof of Theorem 9 proceeds by a direct codimension computation: the line-locus codimension is computed in Equation (8) via an incidence correspondence, and Lemma 10 supplies a lower bound for the codimension of the span-b strata. Lemma 10 is quoted from the same author's published paper [Tse18, Lemma 4.2] with a small modification to the final index condition i_{r-b+a}=k. This is self-citation, but it is not circular: [Tse18] is prior independent published work whose assumptions do not include the all-degrees theorem, and the present note does not define any quantity in terms of the target conclusion. The paper itself flags that the end of the proof of Lemma 10 'can be made more precise' and refers to [Tse18] for the rigorous argument; this is an omitted-proof/rigor gap, not a circular reduction. The numerical comparison following Lemma 10 is carried out in the paper, and no fitted parameter is relabeled as a prediction. No definition encodes the result, no uniqueness claim is imported as an external oracle, and no known pattern is merely renamed. Consequently, there is no significant circularity; the appropriate score is 0.
Assumptions & free parameters
assumptions (4)
- domain assumption Lemma 10 codimension bound for the Span(r,b) stratum, quoted from [Tse18, Lemma 4.2].
- standard math Hilbert function lower bound h_Z(d) >= h_{r,a}(d) for nondegenerate integral subschemes (Lemma 8, from Harris/Park).
- standard math Monotonicity of h_{b,a}(d) in a, used in the chain of inequalities before (11) to replace h_{b,b-i_j+j} by h_{b,1}.
- domain assumption Standard incidence-correspondence dimension counts over an algebraically closed field, including the surjectivity statements around equation (4).
Cite this review
Pith. "Pith review of A note on $r$ hypersurfaces intersecting in $\mathbb{P}^r$." pith.science (2026). https://pith.science/paper/3FQEASNF
@misc{pith2026190801620,
author = {Pith},
title = {Pith review of: A note on $r$ hypersurfaces intersecting in $\mathbbP^r$},
year = {2026},
howpublished = {\url{https://pith.science/paper/3FQEASNF}},
note = {Machine review of arXiv:1908.01620}
}
abstract
We consider the locus of $r$-tuples of homogeneous forms of some fixed degree whose common vanishing locus in $\mathbb{P}^r$ is positive dimensional. We show that any component of maximal dimension of that locus either consists of homogeneous forms all vanishing on some line or homogeneous forms where a proper subset fail to intersect properly.
Reference graph
Works this paper leans on
-
[1]
Joe Harris. Curves in projective space , volume 85 of S\' e minaire de Math\' e matiques Sup\' e rieures [Seminar on Higher Mathematics] . Presses de l'Universit\' e de Montr\' e al, Montreal, Que., 1982. With the collaboration of David Eisenbud
work page 1982
-
[2]
On hypersurfaces containing projective varieties
Euisung Park . On hypersurfaces containing projective varieties. Forum Math. , 27(2):843--875, 2015
work page 2015
-
[3]
The moduli space of hypersurfaces whose singular locus has high dimension
Kaloyan Slavov . The moduli space of hypersurfaces whose singular locus has high dimension. Math. Z. , 279(1-2):139--162, 2015
work page 2015
-
[4]
A note on rational curves on general F ano hypersurfaces
Dennis Tseng. A note on rational curves on general F ano hypersurfaces. Michigan Math Journal . to appear
-
[5]
On degenerate sections of vector bundles
Dennis Tseng. On degenerate sections of vector bundles. preprint . arXiv:1703.10568
-
[6]
Collections of Hypersurfaces Containing a Curve
Dennis Tseng. Collections of Hypersurfaces Containing a Curve . International Mathematics Research Notices , 06 2018
work page 2018
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.