Skip to content

Selection

Origin domain
Mathematics
Subdomain
set and population operations → Mathematics
Also from
Biology & Ecology, Computer Science & Software Engineering, Statistics & Experimental Design, Organizational & Management Science
Aliases
Differential Retention, Selective Retention, Subset Selection
Related primes
Decision, Prioritization, Optimization, Sampling (Representativeness), Selection Bias

Core Idea

Selection transforms an available population by making retention, passage, or influence depend on a criterion, pressure, or rule. Some alternatives survive or count more; others are excluded, suppressed, or assigned less weight. The result is a subset or a population whose composition differs from its input because outcomes were differential rather than uniform.

Selection can be deliberate or impersonal, hard or graded, and one-shot or iterated. A database predicate, a hiring threshold, differential biological reproduction, feature selection, and a membrane admitting some molecules but not others all instantiate the same roles. What matters is not agency but an available population, a selection basis, and unequal continuation.

Structural Signature

  • An eligible population or set of alternatives.
  • A criterion, pressure, predicate, or differential response.
  • Unequal retention, passage, reproduction, inclusion, or weight.
  • A retained output and an excluded or down-weighted complement.
  • A composition change attributable to the selection basis.

Scope of Application

  • Formal systems and databases: select elements satisfying a predicate.
  • Biology and immunity: differential persistence or reproduction of variants.
  • Statistics: draw a sample through an inclusion mechanism.
  • Computing and machine learning: retain features, models, branches, or candidates.
  • Institutions: admit, shortlist, publish, hire, or fund under stated or tacit criteria.
  • Design and maintenance: keep useful components and remove the remainder.

Clarity

Selection distinguishes three neighboring moves. Decision commits to one path; selection may retain many alternatives and need not involve an agent. Prioritization orders alternatives; selection changes which alternatives continue or how strongly they count. Optimization seeks a best feasible configuration; selection can be stochastic, biased, local, or merely thresholded and need not find an optimum.

Abstract Reasoning

The abstraction directs attention to the unobserved complement. Whenever only the survivors are visible, ask what population entered, what rule or pressure operated, which alternatives did not pass, and how conditioning on passage changed the output. Changing the criterion changes the population even when the candidates do not change.

Example

A system starts with alternatives (A) and a predicate ©. Hard selection returns (S_c(A)={x\in A:c(x)=1}). Graded selection assigns weights (w(x)), so alternatives contribute in different proportions without a binary cutoff. A database query, an evolutionary population, and an admissions process can all be mapped onto this same operator.

Relationships to Other Abstractions

Local relationship map for SelectionParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.SelectionPRIMEDomain-specific abstraction: Uses and Gratifications — is part ofUses andGratificationsDOMAINDomain-specific abstraction: Market Pull — is a decomposition ofMarket PullDOMAINDomain-specific abstraction: Precedence Effect — is a decomposition ofPrecedenceEffectDOMAINDomain-specific abstraction: Snowball Sampling — is a kind ofSnowballSamplingDOMAINPrime abstraction: Adverse Selection — is a kind ofAdverseSelectionPRIMEPrime abstraction: Gatekeeping — is a kind ofGatekeepingPRIMEPrime abstraction: Gresham's Law — is a kind ofGresham's LawPRIMEPrime abstraction: Natural Selection — is a kind ofNaturalSelectionPRIMEPrime abstraction: Pruning — is a kind ofPruningPRIMEPrime abstraction: Structural Filtering — is a kind ofStructuralFilteringPRIMEPrime abstraction: Theoretical Sampling — is a kind ofTheoreticalSamplingPRIME

Current abstraction Selection Prime

Foundational — no parent edges in the catalog.

Children (11) — more specific cases that build on this

  • Snowball Sampling Domain-specific is a kind of Selection

    Snowball sampling is selection specialized to unequal sample inclusion generated by reachability from seeds through participant referral ties.

  • Adverse Selection Prime is a kind of Selection

    Adverse Selection is selection specialized to hidden pre-contractual types whose unequal willingness to accept uniform terms degrades the participant pool.

  • Gatekeeping Prime is a kind of Selection

    Gatekeeping is selection specialized to a choke point where an arbiter controls passage and the downstream audience cannot directly observe the rejected stream.

  • Gresham's Law Prime is a kind of Selection

    Gresham's Law is selection specialized to an enforced nominal parity that gives lower-quality goods greater passage through a channel and higher-quality goods greater retention outside it.

  • Natural Selection Prime is a kind of Selection

    Natural selection is selection specialized to heritable variants whose unequal reproduction or persistence shifts population composition across repeated rounds.

Not to Be Confused With

  • Natural Selection adds heritable variants, differential reproduction or persistence, and iteration across rounds.
  • Pruning adds a prior over-generation phase and removes components from an already built system.
  • Gatekeeping adds a choke point, an arbiter, discretion, asymmetric power, and an unobserved rejected stream.
  • Structural Filtering adds several parallel institutional filters whose intersection shapes output without a focal discretionary gatekeeper.
  • Screening is the economic mechanism in which an uninformed designer offers a menu that induces informed agents to reveal type through self-selection.
  • Sampling Representativeness evaluates whether a statistical selection procedure supports inference to a target population.

Notes

(New generic intermediate exposed by the recursive mixed-DAG pass; queued for Claude style and citation review.)