My second qualifying paper is on transphonologization of the LowBackVowels in CaliforniaEnglish.
Vowel identification
One of the things we know when we learn a language is the way units can be put together. At a certain age, a child is able to understand that some morphemes are productive, and can begin to produce novel utterances by combining different morphemes to arrive at the intended meaning. In the case of morphology we can show quite clearly that language is organized as discrete units that can be recursively combined. This phenomenon extends to other aspects of language production from the organization of whole words within sentences (Syntax) and down to the fine-grain phonetic categorization of sounds (Phonology). Speech as a physical event however is not a discrete (digital) process but is a continuous (analog) process in the production of acoustics.
- An Ende example would be nice here, it can be any generic sentence
- .-.. .- -. --. ..- .- --. .
In order for speaker A to produce the sentence in (1) for speaker B a number of steps must be accomplished. Speaker A must envision some idea. To convey this idea, speaker A must decide what sequence of discrete lexical items is able to convey her idea from a list of items she expects speaker B to know and be able to understand and act upon if necessary. This step we will call the EncodingStep whereby speaker A encodes a message which she expects speaker B to be able to decode. After the message is encoded speaker A must execute a sequence of muscle movements which produce time-varying signals through some medium. For oral-aural languages, this is a time-varying acoustic signal caused by the disruption of an airstream. For manual-visual languages, it is a time varying visual signal caused by movement in space transmitted by light to the retina. This step of producing a time-varying signal we will call the ProductionStep. The EncodingStep and the ProductionStep both rely on the language-specific set of LexicalItems in the lexicon. The lexicon defines the code which is used in the EncodingStep. For the encoding step to succeed, the lexicon must provide it with what a lexical item means (Semantics; sign and significant) and the way the item can be arranged in relation to other lexical items (Syntax, Morphology, Phonology). For the ProductionStep to succeed the LexicalItem must provide the particular sequence of muscle movements required to produce the signal recognized as the sign which signifies the intended meaning encoded in the message (Phonetics-Phonology interface).
Imagine now that speaker A transmitted the sentence in (2) produced in a language with two phonemes, /./ and /-/. Speaker B must perceive the signal produced in the ProductionStep through one or more of his senses. We will call this process of converting an analog signal to a digital signal the PerceptionStep. Once the signal has been divided up into its sequence of LexicalItems, speaker B is able to recover the meaning intended by speaker A in the EncodingStep by associating particular sequences of lexical items with their conventionalized significant. We will call this process of recovering a message from a digital signal the DecodingStep. For the PerceptionStep to succeed, the LexicalItem must specify what particular analog signals correspond to it. For oral languages, this association is primarily specified in relation to the acoustic signal, while for manual languages it is the visual or tactile signal. For the DecodingStep to succeed, the LexicalItem must specify what meanings are associated with a particular digital signal.
To summarize:
These steps are performed in parallel, rather than sequentially, by making predictions and sharing partial answers with each other. We can then understand these steps better as functions. The input of each function is the output of the other functions, and a step is complete when its outputs stabilize. Early on in the utterance, there is limited information and so predictions about the ultimate signal are highly variable, all else being equal. As more of the utterance is completed, there is more information to be processed leading to better informed and more consistent outputs.
TransPhonologization
Larry Hyman coins the term TransPhonologization in his chapter in Yu 2013. It is a form of grammaticalization whereby a different feature becomes phonologized as the new contrastive marker.