Taxonomy for Humans or Computers? Cognitive Pragmatics for Big Data

Research output: Contribution to journalArticlepeer-review

13 Scopus citations

Abstract

Criticism of big data has focused on showing that more is not necessarily better, in the sense that data may lose their value when taken out of context and aggregated together. The next step is to incorporate an awareness of pitfalls for aggregation into the design of data infrastructure and institutions. A common strategy minimizes aggregation errors by increasing the precision of our conventions for identifying and classifying data. As a counterpoint, we argue that there are pragmatic trade-offs between precision and ambiguity that are key to designing effective solutions for generating big data about biodiversity. We focus on the importance of theory-dependence as a source of ambiguity in taxonomic nomenclature and hence a persistent challenge for implementing a single, long-term solution to storing and accessing meaningful sets of biological specimens. We argue that ambiguity does have a positive role to play in scientific progress as a tool for efficiently symbolizing multiple aspects of taxa and mediating between conflicting hypotheses about their nature. Pursuing a deeper understanding of the trade-offs and synthesis of precision and ambiguity as virtues of scientific language and communication systems then offers a productive next step for realizing sound, big biodiversity data services.

Original languageEnglish (US)
Pages (from-to)99-111
Number of pages13
JournalBiological Theory
Volume12
Issue number2
DOIs
StatePublished - Jun 2017
Externally publishedYes

Keywords

  • Big data
  • Cognitive pragmatics
  • Concept taxonomy
  • Data aggregation
  • Knowledge representation and reasoning
  • Nomenclature

ASJC Scopus subject areas

  • Ecology, Evolution, Behavior and Systematics
  • History and Philosophy of Science

Fingerprint

Dive into the research topics of 'Taxonomy for Humans or Computers? Cognitive Pragmatics for Big Data'. Together they form a unique fingerprint.

Cite this