Linguist Forum

Specializations => Phonetics and Phonology => Topic started by: Copernicus on June 07, 2015, 12:40:48 PM

Title: What is Natural Phonology?
Post by: Copernicus on June 07, 2015, 12:40:48 PM
I've already declared my theoretical bias as a proponent of David Stampe's theory of Natural Phonology  in the Phonology is not grounded in phonetics ( thread.  Rather than for me to go into my own description of the theory here, I'd rather just refer people to a published (2009) paper called Hypotheses of Natural Phonology ( by David Stampe and Pat Donegan.   It is roughly 26 pages of reading.  This paper provides a summary of the basic ideas and relates it to mainstream phonology, especially the popular Optimality Theory that seems to be dominant today.

I was a student of David's at Ohio State in the 1960s and 1970s, when he first proposed the theory.  Since then, it hasn't really caught on anywhere except in a few places in Europe.  Natural Phonology used to be more well-known than it is today, but I've met younger linguists who have never heard of it.  It is frequently confused with Joan Bybee's similar-sounding "Natural Generative Phonology", but the two theories are not related at all.  Joan's theory is a generative theory of phonology that attempts to establish a strong boundary between so-called low level or automatic rules and high level or morphophonemic rules.  Natural phonology maintains a similar dichotomy, but it is not a generative theory.  That is, Natural Phonology purports to be directly relevant to speech production and perception.  Although I don't think David or Pat have said much in print about their views on well-formedness intuitions, a basis of generative theory, I would argue that well-formedness intuitions are derivative of performance strategies up to a point.  That is, they derive from our general cognitive ability of introspective empathy.  We can imagine how we or someone else might produce a linguistic expression.
Title: Re: What is Natural Phonology?
Post by: Copernicus on June 08, 2015, 12:31:49 PM
If you read the "Hypotheses" paper that was delivered in Poznan in 2009, you'll note that there are several criticisms of OT along with the areas of overlap between the theories.  I personally think that the biggest problem with OT is that it still adheres to basic generative assumption that grammars exist primarily to validate structural well-formedness.  However, it does share one very radical notion with Natural Phonology (NatP) that did not exist in classical generative theory--the lack of any constraints on inputs, so-called "richness of the base". 

From a NatP perspective, I would just put it this way--people can try to articulate any phonetic target they can imagine.  That is, one can try to say "bnick", even though the initial cluster is clearly ill-formed for English.  NatP proposes that there are two basic types of phonological processes:  fortitions and lenitions.  Fortitions (like CON?) are clarifying processes that tend to make different sounds more salient and less like each other.  So epenthesis would be a fortition.  Lenitions (like EVAL?) are simplifying processes that make sounds more like each other in context (e.g. assimilation) and tend to reduce structure.  So syncope would be a lenition in that it reduces the number of syllables produced during execution.  Unlike OT, however, NatP allows easily for processes to produce rather opaque forms.  So voicing assimilation can be "fed" by syncope, and voice assimilation might "feed" other lenitions.  So you can get extremely radical deviations from input in fast and casual speech.  (Stampe once wrote a short paper called "Divinity Fudge" published by the Chicago Linguistic Society that went into detail on the wide range of casual pronunciations of that phrase.)

The theory of Natural Phonology holds that infants start out with roughly adult phonemic forms even when the child produces extremely deviant outputs in relation to the underlying forms.  Hence, NatP is able to explain the stages that children go through as their process system is gradually brought under control.  OT has nothing much to say about speech outputs in the earliest stages of language production.