Young Functions: Analysis & Combinatorics

• Young functions are convex growth functions and measurable mappings used in Orlicz spaces and to capture oscillatory behavior in analysis.
• They facilitate generalized proximal regularization in uniformly convex Banach spaces by ensuring uniform continuity via strict convexity.
• In combinatorics, Young functions denote basis elements on Young diagrams and quasisymmetric Schur functions, offering positive expansion properties.

“Young functions” appears in several distinct mathematical settings. In Orlicz-space theory it denotes a convex function $\Psi:[0,\infty)\to[0,\infty)$ used to define Orlicz spaces and the Luxembourg norm; in uniformly convex Banach spaces it furnishes generalized proximal regularization; in recent geometric measure theory it denotes a measurable map into a space of probability Radon measures; and in algebraic combinatorics closely related phrases refer either to polynomial functions on Young diagrams or, in the terminology of some quasisymmetric-function papers, to Young quasisymmetric Schur functions (Rodney et al., 2022, Bacak et al., 2017, Chou, 7 Oct 2025, Aval et al., 2013, Allen et al., 2016).

1. Terminological scope

The shared name masks non-equivalent objects. In the Orlicz setting, a Young function is a scalar growth function on $[0,\infty)$; in the varifold setting, a Young function is a measurable map $f:X\to \mathbf{P}(Y)$ taking values in probability measures; in algebraic combinatorics, one encounters functions on Young diagrams defined through interlacing or multirectangular coordinates, and also Young quasisymmetric Schur functions indexed by compositions (Rodney et al., 2022, Chou, 7 Oct 2025, Aval et al., 2013, Allen et al., 2016).

A common source of confusion is the proximity of these notions to Young measures. The measure-theoretic literature represented here uses “Young measure” for parametrized probability measures associated with oscillatory or Borel functions, while the 2025 varifold paper uses “Young function” for a measurable probability-valued map itself. The combinatorial literature, by contrast, attaches the adjective “Young” to diagrams, tableaux, and composition tableaux rather than to convex growth functions (Grzybowski et al., 2016, Puchała, 2018, Chou, 7 Oct 2025).

2. Young functions in Orlicz analysis

In the Orlicz-space setting, a Young function is a function of the form

$\Psi(t)=\int_0^t \psi(s)\,ds,$

where $\psi$ is right-continuous, non-decreasing, $\psi(0)=0$, and $\psi(t)>0$ for $t>0$. The resulting $\Psi$ is continuous, strictly increasing, convex, satisfies $\Psi(0)=0$, and obeys $[0,\infty)$0 as $[0,\infty)$1 (Rodney et al., 2022).

Given a measure space $[0,\infty)$2, the associated Orlicz space $[0,\infty)$3 consists of measurable $[0,\infty)$4 such that the Luxembourg norm

$[0,\infty)$5

is finite. The special choice $[0,\infty)$6 recovers $[0,\infty)$7. The paper also records a Chebyshev-type inequality,

$[0,\infty)$8

and the exact norm of indicators,

$[0,\infty)$9

which make the threshold behavior of $f:X\to \mathbf{P}(Y)$0 directly visible in norm estimates (Rodney et al., 2022).

These formulas are central because later asymptotic results are expressed in terms of families $f:X\to \mathbf{P}(Y)$1 and the convergence of $f:X\to \mathbf{P}(Y)$2. The framework is entirely norm-theoretic: bounded functions, indicators, and inverse Young functions determine the limit behavior.

3. Families, asymptotic limits, and proximal regularization

For one-parameter families $f:X\to \mathbf{P}(Y)$3, the key notion is $f:X\to \mathbf{P}(Y)$4-admissibility. A family is $f:X\to \mathbf{P}(Y)$5-admissible if $f:X\to \mathbf{P}(Y)$6 for $f:X\to \mathbf{P}(Y)$7, while for $f:X\to \mathbf{P}(Y)$8 one has either $f:X\to \mathbf{P}(Y)$9 when $\Psi(t)=\int_0^t \psi(s)\,ds,$0, or $\Psi(t)=\int_0^t \psi(s)\,ds,$1 when $\Psi(t)=\int_0^t \psi(s)\,ds,$2. Proposition 2.5 gives the equivalent inverse formulation

$\Psi(t)=\int_0^t \psi(s)\,ds,$3

Under the monotonicity hypothesis that $\Psi(t)=\int_0^t \psi(s)\,ds,$4 is non-decreasing on $\Psi(t)=\int_0^t \psi(s)\,ds,$5 for all $\Psi(t)=\int_0^t \psi(s)\,ds,$6 and large $\Psi(t)=\int_0^t \psi(s)\,ds,$7, the main theorem states

$\Psi(t)=\int_0^t \psi(s)\,ds,$8

for all $\Psi(t)=\int_0^t \psi(s)\,ds,$9 if and only if $\psi$0 is $\psi$1-admissible. The examples include powers $\psi$2, log-bumps $\psi$3, iterated log-bumps, and compositions with another Young function; the counterexamples include $\psi$4 on $\psi$5 with Lebesgue measure and further non-$\psi$6-admissible oscillatory families (Rodney et al., 2022).

In uniformly convex Banach spaces, the proximal-mapping paper imposes a stronger analytic hypothesis: a Young function $\psi$7 is continuous and strictly convex, with

$\psi$8

It defines the generalized proximal mapping

$\psi$9

For $\psi(0)=0$0, this recovers the usual $\psi(0)=0$1-proximal mapping. A central quantitative result is that $\psi(0)=0$2 is uniformly convex on bounded subsets, with an explicit modulus

$\psi(0)=0$3

where $\psi(0)=0$4 and $\psi(0)=0$5 are given in terms of moduli $\psi(0)=0$6, $\psi(0)=0$7, $\psi(0)=0$8, and $\psi(0)=0$9. This leads to uniform continuity of $\psi(t)>0$0 on bounded sets, with explicit modulus

$\psi(t)>0$1

The paper also distinguishes this scaling from the alternative regularizer $\psi(t)>0$2, noting that the chosen form avoids degeneration of the modulus of uniform continuity as $\psi(t)>0$3 (Bacak et al., 2017).

4. Measure-valued Young functions and their relation to Young measures

In the 2025 varifold paper, a Young function is a measure-theoretic model for multiple-valued functions. If $\psi(t)>0$4 and $\psi(t)>0$5 are locally compact Hausdorff spaces and $\psi(t)>0$6 is a Radon measure on $\psi(t)>0$7, then a $\psi(t)>0$8-Young function of type $\psi(t)>0$9 is a $t>0$0-measurable map

$t>0$1

where $t>0$2 denotes probability Radon measures on $t>0$3 with the weak topology. Ordinary functions $t>0$4 embed as Dirac-valued Young functions $t>0$5. The associated graph measure is

$t>0$6

and convergence of pairs $t>0$7 is defined as narrow convergence of the graph measures $t>0$8. The paper proves compactness criteria for sets of graph measures, a disintegration theorem $t>0$9, and a compactness theorem for convergent pairs of varifolds and Young functions under tightness in the codomain variable. It also introduces codomain test-function spaces $\Psi$0, $\Psi$1, product spaces such as $\Psi$2, and a pseudo-metric

$\Psi$3

which agrees with the $\Psi$4-Wasserstein metric on $\Psi$5 (Chou, 7 Oct 2025).

Related Young-measure papers clarify the surrounding measure-theoretic background. For a Borel function $\Psi$6, the Young measure $\Psi$7 is the probability distribution of $\Psi$8, where $\Psi$9 is uniformly distributed on $\Psi(0)=0$0; equivalently,

$\Psi(0)=0$1

Simple Young measures are weak* dense among Young measures associated with measurable functions (Grzybowski et al., 2016). For homogeneous Young measures associated with $\Psi(0)=0$2-oscillating functions, the density is

$\Psi(0)=0$3

and weak convergence follows when the total slopes form a monotonic sequence (Puchała, 2018). This suggests that the varifold construction belongs to a broader measure-theoretic framework in which probability-valued objects encode oscillation, multiplicity, and weak limits.

5. Polynomial functions on Young diagrams

A different use of the phrase concerns functions defined on Young diagrams. The paper “Quasi-symmetric functions as polynomial functions on Young diagrams” studies polynomials in infinitely many interlacing coordinates $\Psi(0)=0$4 whose value depends only on the Young-diagram shape. The defining condition is the stability relation

$\Psi(0)=0$5

which expresses that coinciding consecutive interlacing coordinates correspond to the same diagram. The space $\Psi(0)=0$6 of such stable polynomials is an algebra isomorphic to $\Psi(0)=0$7. The isomorphism is realized by evaluating quasi-symmetric functions on the virtual alphabet

$\Psi(0)=0$8

so that $\Psi(0)=0$9. The monomial quasi-symmetric functions $[0,\infty)$00, indexed by compositions $[0,\infty)$01, form a basis of $[0,\infty)$02; in homogeneous degree $[0,\infty)$03, the dimension is $[0,\infty)$04, or $[0,\infty)$05 after factoring out the alternating-sum relation $[0,\infty)$06. The same algebra is obtained from multirectangular coordinates $[0,\infty)$07 by imposing the rectangular-invariance relations displayed in the paper, and there is also a noncommutative analogue isomorphic to $[0,\infty)$08 (Aval et al., 2013).

This description sits alongside two related lines of work. First, the Kerov–Olshanski algebra $[0,\infty)$09 of polynomial functions on Young diagrams includes normalized characters, free cumulants, and the functionals

$[0,\infty)$10

and a graph-theoretic criterion is given for when embedding counts $[0,\infty)$11 are polynomial in this sense: $[0,\infty)$12 (Dołega et al., 2011). Second, shifted Schur functions $[0,\infty)$13 and the family $[0,\infty)$14 are polynomial in multirectangular coordinates and have nonnegative coefficients when expressed in the falling-factorial basis; the paper further proposes a Jack-theoretic extension and proves it for one-part partitions (Alexandersson et al., 2016).

6. Young quasisymmetric Schur functions and peak analogues

In the terminology of one quasisymmetric-function paper, “Young functions” refers to Young quasisymmetric Schur functions. For a composition $[0,\infty)$15, the function

$[0,\infty)$16

is defined by summing over semistandard Young composition tableaux $[0,\infty)$17 of shape $[0,\infty)$18. The same paper proves that dual immaculate quasisymmetric functions decompose positively in this basis: $[0,\infty)$19 where $[0,\infty)$20 is the number of dual immaculate recording tableaux (DIRTs) of shape $[0,\infty)$21 and row strip shape $[0,\infty)$22. The proof uses a Schensted-type insertion and an inverse “rapture” procedure. It also gives a symmetry criterion: a dual immaculate function is symmetric if and only if $[0,\infty)$23 for some $[0,\infty)$24; in that case it agrees with the corresponding hook Schur function (Allen et al., 2016).

Two 2024 papers develop peak analogues. One introduces a bijection

$[0,\infty)$25

on standard peak immaculate tableaux such that

$[0,\infty)$26

and hence the descent distributions for column and row reading words coincide. Using this equidistribution together with the Allen–Hallam–Mason insertion algorithm, it proves that the transition matrix from quasisymmetric Schur $[0,\infty)$27-functions $[0,\infty)$28 to peak Young quasisymmetric Schur functions $[0,\infty)$29 is upper triangular with non-negative integer coefficients with respect to dominance order: $[0,\infty)$30 (Choi et al., 2024). The second paper proves the corresponding positive expansion in the peak algebra,

$[0,\infty)$31

where $[0,\infty)$32 counts standard peak Young composition tableaux or, equivalently, DIRTs with row strip shape $[0,\infty)$33. It also identifies the single-term cases: $[0,\infty)$34 (Searles et al., 2024).

Taken together, these combinatorial results show that “Young functions” in this strand of the literature are basis elements in quasisymmetric or peak algebras, defined through composition tableaux and related insertion procedures rather than through convexity or measure-valued maps. The term therefore remains fundamentally context-dependent.