Realizing that many linguists, young and old, find themselves unsure of how best to succeed (or have success thrust upon them), we of the Speculative Grammarian Editorial Board have assembled a collection of high-
You can represent some of the patterns simply all of the time, or all of the patterns simply some of the time, but you can’t represent all of the patterns simply all of the time.
Control definitions, and you control the means of model-
The smart way to keep linguists passive and obedient is to strictly limit the spectrum of acceptable theories, but allow very lively debate within that spectrum.
Go not to the linguists for usage rules, for they shall say both yes and no
Do not meddle in the affairs of physicists, for they are subtle and quick to oversimplify.
Those who fail to learn linguistic history are doomed to recreate it in their synchronic models.
Never try to teach a syntactician to address sociolinguistic context; it wastes your time and it annoys the syntactician.
The higher the number of people who speak the language you are studying, the greater the number of errors in your data.
No one ever got rich betting they could find something new in English linguistics.
If you can’t write, study literature. If you can’t read either, study literary linguistics.
One does not simply walk into fieldwork.
If you try to study translation without knowing two languages and doing the job first, you’re gonna have a baaad time.
Never underestimate the explanatory power of a single Hopi speaker in New York City!
Keep re-processing that fMRI data, you’ll find correlates sooner or later!
When collecting phonetic data, remember that speakers can only be disqualified or excluded when (# participants) > (desired # participants).
When collecting reaction time data from undergraduates, the pervasive vodka-
Any article or chapter not available in digitized form probably wasn’t worth citing anyways.
Any data that questions your work was gathered wrongly.
And any work that questions your data is disingenuous.