Chapter 5
Are new words predictable?
A pilot study on the origin of neologies by means of natural
selection
Research on neologisms has so far been dominated by
retrospective investigations (cf. e.g. Schmid 2008; Kerremans & Prokić 2018). By contrast,
the present study is probably the first to directly tap into speakers’
linguistic competence by eliciting names for unfamiliar entities.
Building on early assumptions by Darwin (1859, 1871) and others an evolutionary, fitness-based approach
is developed to predict und thus explain the selections made by groups
of German speakers of successful neologisms from self-generated pools.
Four main factors are identified as crucial: quantity (brevity),
quality (truthfulness), relation (distinctiveness) and manner
(attractiveness). In order to test the viability of the theory an
experiment is conducted and replicated. The results provide
preliminary support for the assumptions made.
Article outline
- 1.Introduction
- 1.1Predictive language theories and their moving target
- 1.2Language change and neology
- 1.3Language evolution: From Darwin via Haeckel, Schleicher and Müller back to
Darwin
- 1.4The incredible career of a neologism and Engelbart’s
Question
- 1.5Naming and categorization: Clarifications and sample case observations
- 2.Investigating form and fate of neologisms: An experimental approach
- 2.1Creating demand for names of unknown objects: Meet the Fribbles
- 2.2Preparatory data collection: Five objects, five name pools and an initial
hypothesis
- 2.3Venturing a first hypothesis
- 3.Pretest: Deriving preference hypotheses from estimated natural
selection
- 3.1Pretest
- 3.2Pretest results: Five pools resulting from artificial and estimated natural
selection
- 3.3Developing a selection predictor: Factors of fitness ranking
- 1.Quantity
- 2.Quality
- 3.Relation
- 4.Manner
- 3.4Assigning degrees of fitness: The coding algorithm
- 1.Quantity
- 2.Quality
- 3.Relation
- 4.Manner
- 4.Experiment: Testing the fitness rank predictor with forced natural
selection
- 4.1Experiment
- 4.2Comparing observed and predicted rankings
- 4.2.1Object 1 experiment ranking
- 4.2.2Object 2 experiment ranking
- 4.2.3Object 3 experiment ranking
- 4.2.4Object 4 experiment ranking
- 4.2.5Object 5 experiment ranking
- 4.2.6Overview of experiment rankings
- 4.3Discussion of the experiment results
- 4.3.1Limits of ecological validity
- 4.3.2Other possible constraints
- 4.3.3Tackling the replicability issue
- 5.Replication: Retesting the fitness rank predictor
- 5.1Replication procedure
- 5.2Comparing observed and predicted replication rankings
- 5.2.1Object 1 replication ranking
- 5.2.2Object 2 replication ranking
- 5.2.3Object 3 replication ranking
- 5.2.4Object 4 replication ranking
- 5.2.5Object 5 replication ranking
- 5.2.6Overview of replication rankings
- 5.3Discussion of the replication
- 5.3.1Group dynamics
- 5.3.2Optimizing fitness categories?
- 6.Overall analysis
- 6.1Ways of estimating correlation strength of incomplete rankings
with ties
- 6.2Statistical overview of experiment and replication
results
- 7.Summary and outlook
-
Acknowledgements
-
Notes
-
References