Aspiration in New Julfa Armenian
A couple of days ago, I attended a very interesting talk by Sebouh Aslanian on New Julfa Armenians and their trade networks, stretching from Amsterdam to Manila in the Philippines. New Julfa is an Armenian settlement in Isfahan, Iran, located along the south bank of the Zayandeh River. Its historical roots go back to the city of Julfa (also known as Jugha or Djugha) in Nakhichevan (present-day Azerbaijan). In 1606 Shah Abbas I, an influential shah from the Safavid dynasty moved some 150,000 Armenians from Julfa to Isfahan, founding the New Julfa community. This move was due to forced resettlement by Abbas I himself (though some Iranian accounts disagree). What the various scholars do agree on is that New Julfa Armenians were adventurous traders in silk. They established trusted relationships with members of their own community who have spread over the silk trade roots. These trade neworks are discussed in more detail in a GeoCurrents‘ post by Martin W. Lewis.
But New Julfa Armenians are interesting not only from the point of view of their trade networks, but for their dialect too. As is typical for many emigre communities, New Julfa Armenians preserved some features of the ancestral tongue better than other dialects remaining in the homeland. One aspect of the New Julfa Armenian that is of interest to linguists is that it allows us to examine the issue of aspiration in fricatives.
Aspiration refers to a short (about 25 milliseconds) delay in the spread of the vocal cords after a voiceless stop is released. For example, consider the articulation of the word pot in English. The first segment [p] is voiceless, so the vocal cords are spread to let the air pass through the glottis unimpeded. As for the oral tract, the lips are first closed and then released. The second segment, the vowel, is voiced so the vocal cords are brought close together to create vibration. In the oral cavity, the vowel is articulated as low, back, unrounded, and lax. Aspiration results from the non-alignment of the laryngeal (vocal cords) and oral (mouth) articulations: when the mouth is already pronouncing the vowel, the vocal cords remain for a brief moment in the voiceless/spread position.
In English aspiration is non-phonemic, as its appearance is predictable from context. Specifically, voiceless stops are aspirated in stressed syllable-initial position, for example in tack and attack, but not in stack. But the same is not true in several other languages, where aspiration serves the distinctive function (that is, it can distinguish one word from another). For example, in Thai ta: (without aspiration) means ‘eye’ and tha: (with aspiration) means ‘landing place’. Similarly, in Hindi kal (without aspiration) means ‘era’ and khal (with aspiration) means ‘skin’. Armenian too uses aspiration distinctively: kap (without aspiration) means ‘bond’ and kaph (with aspiration) means ‘club’; mut (without aspiration) means ‘entrance’ and muth (with aspiration) means ‘darkness’; and tak (without aspiration) means ‘under’ and takh (with aspiration) means ‘hot’.
Phonetically speaking, aspiration most typically applies to stops (though some languages, such as Burmese, have aspirated fricatives too). But what about the mental representation of fricatives? Are they grouped into the same natural class with aspirated or unaspirated stops? The New Julfa Armenian dialect sheds some light on this issue. As discussed in a paper by Bert Vaux “The Laryngeal Specifications of Fricatives”, New Julfa Armenian has a four-way (2×2) contrast in its inventory of stops: voiced unaspirated, voiced aspirated, voiceless unaspirated, and voiceless aspirated. Fricatives, on the other hand, only come in voiced and voiceless varieties, and never carry phonetic aspiration. But a certain assimilation phenomenon reveals that voiced and voiceless fricatives are categorized differently in this dialect: voiced fricatives pattern with unaspirated stops and voiceless fricatives with aspirated stops.
The assimilation in question is a morphophonological rule operating on the future morpheme, the prefix k-. This prefix assimilates in voicing and aspiration to the following consonant; the basic form k- is retailed if it attaches to a stem that starts with a vowel. The prefix remains k- if it attaches to a stem that starts with an unaspirated voiceless consonant, such as the stem tam ‘go’ (an epenthetic schwa is inserted between the prefix and the stem to allow for syllabification, but this insertion must apply after the assimilation rule; we don’t need to concern ourselves with this detail though). When the future tense prefix is attached to a stem that starts with a voiced unaspirated consonant, e.g. lam ‘cry’, it is voiced, turning into a g-. When the prefix is attached to a stem that begins with an aspirated voiceless stop such as thorniem ‘allow’, the resulting form of the prefix is kh-, and when it attaches to a stem beginning with an aspirated voiced stop such as ghom ‘come’, the resulting form of the prefix is gh-. What happens when the future tense prefix attaches to a stem that starts with a fricative, such as savoriem ‘grow accustomed to’ or zram ‘bray’? In the latter case, the prefix becomes g-, that is it assimilates in voicing but there is no aspiration added. But in the case of stems starting with a voiceless fricative, like savoriem ‘grow accustomed to’, the resulting form of the prefix is kh-, meaning that voiceless fricatives are mentally treated not only as voiceless but as aspirated as well. The image below summarizes these data from Vaux’s paper.
Like this post? Please pass it on:
« Syntactic feature or scribal convention?
“Oh, rather!” »