Julia Trippe – trippe@uoregon.edu
Eric Pederson
University of Oregon
Linguistics Department
Eugene, OR 97403-1290
Popular version of paper 5pSC32
Presented Friday afternoon, December 6, 2013
166th ASA Meeting, San Francisco
In 1951, the International Civil Aviation Organisation (ICAO) declared English the official language to be used between international pilots and controllers. Since that time there have been several major accidents and many incidents related to lack of clear communication between controllers and pilots (2004 ICAO Assembly). This state of affairs prompted the ICAO to require English training and testing for all international commercial pilots starting in 2011. But, even today, pilots are meeting only the minimum requirements of Aviation English (AE) proficiency (Kim and Elder, 2011).
AE is not standard conversational English, it is a sub-variety of English: a codified, abbreviated, jargon-filled form of English, which must be learned even by Native English speakers. Generally this process is undertaken during the course of flight training and can be one of the most daunting aspects of learning to fly. For non-native English speakers, the process is even more difficult. Many flight students come to the US to train because of the availability of small aircraft and the training infrastructure. Although they may have had several years of classroom English in their home country, these students are usually not yet comfortable conversing in Standard English. Therefore, their learning curve is steeper than for native English students. Learning to control an aircraft while trying to decipher the rapid-fire jargon of American pilots and controllers is not the right environment in which to learn this new language. The question arises: how and where and to what standard should we be training these pilots?
Previous studies measure AE speech accuracy by task performance and number of correctly repeated elements (Barshi and Healy 2011) and speech comprehensibility using native speaker judgments (Farris et al 2008). The current study develops a quantitative index for evaluating AE production based on rhythm. We are developing a profile of Native Speaker AE with which to evaluate Non-native Speaker AE production. This profile will ultimately help guide training methods for pilots learning AE coming from different first language (L1) backgrounds.
There are many challenges to communicating in the Aviation English environment. Primarily, there is a lack of contextual cues (facial expression, body language, shared physical space, etc.). Another important cue that is missing in AE is intonation, because of the restricted bandwidth of radio transmissions (4000 Hz, similar to a cellular telephone). This limitation also makes some sounds ambiguous (i.e. s, th and f). With the addition of radio static and other speakers on the same frequency, AE can be unclear in the best of circumstances. Other factors contributing to lack of clarity in the AE speech signal are the use of jargon and lack of disambiguating function words (a, the, of, for, etc.) as well as the usually rapid speech rate. When all of these factors are accounted for, the one reliable source of information in the speech signal is rhythm. If non-native AE users are to be safe and reliable users of the global air traffic system, they need to be able to understand and produce native-like AE rhythm patterns.
The search for the most salient aspect of comprehensibility in language leads us to examine child language acquisition. Prosody (rhythm and intonation of language) is one of the first things humans learn in their native language (Ramus 2006). Newborn infants recognize the rhythm of their mother's language (Nazzi, et al 1998) and cry with the same pitch and rhythm patterns of their native tongue (Mampe et al 2009). Prosody is also one of the more difficult aspects of learning a second language (Guion 2005). Advanced adult learners of second languages - people who conduct their daily affairs in a language other than their first, persist, even after many years, in using the rhythm of their native tongue. This non-native speech timing is much of what native speakers of a language perceive as 'foreign accent', which is directly correlated to lack of comprehensibility (Tajima et al 1997).
Phoneticians have looked for a discrete acoustic feature, something discernable in the sound waves themselves that is captured by human perception of rhythm. Over the years, we have gone from describing what we perceive as stress vs. syllable timed languages (Abercrombie 1967) to recognizing that there are other key elements in speech timing that create a recognizable profile. In an effort to capture the timing differences that a child's ear catches naturally, linguists have derived measures that focus on vowel and/or consonant durations. Much of this research has focused on the variability in individual vowel and consonant durations. Stress-timed languages are those, like English, that can reduce the length of a vowel in a non-stressed syllable and have a wide variety of possible consonant structures (up to four consonants in a cluster is not uncommon). On the other hand, Syllable-timed languages like Korean do not reduce vowel length and have severely restricted consonant usage (typically one per syllable).
Since AE is different from Standard English in that it does not utilize short, typically unstressed/reduced function words, we expect AE to behave more like a syllable-timed language than Standard English does. But, since AE still allows the complex consonant structures represented in Standard English, we expect it to exhibit different characteristics than syllable-timed languages.
The first stage of our research is to establish an index that describes the rhythm of AE. To this end, we have chosen to use productions of native English speaking Air Traffic Controllers practicing in the US (Godfrey and Linguistic Data Consortium 1994). By using only native speaker American professionals, we hoped to capture a target-like standard that can be compared to measurements from other languages. Dialog was extracted from tapes of ATC stations throughout the US. Utterance-length transmissions were then paired with their transcriptions in an automatic speech aligner that divided the speech stream into segments of sound (Gorman, et al 2011). After checking and cleaning up the data, we ran a program to extract the vowels and consonants and generate comparisons following each of the metrics we wished to capture.
We then ran Standard American English speech through the same aligner, so that we could equitably compare them. This process will give us an idea of how different or similar AE is from Standard American English. We will also compare our measurements to existing measures from other scholars. If our measures are comparable to theirs, we can begin to compare how AE fits in the overall picture of language rhythm.
The goal of this research is to show that Aviation English exhibits a distinct rhythmic signature, different from Standard English, yet consistent and quantifiable in itself. In the future, we intend to measure the rhythm of non-native AE speakers as well as native student AE speakers for comparison. These measures are primarily meant as an assessment tool by which individuals can be categorized in their Aviation English proficiency, problem areas can be identified and progress can be measured.
References:
Abercrombie, David. (1967). Elements of general phonetics (Vol. 203): Edinburgh University Press Edinburgh.
Barshi, Immanuel, and Healy, Alice F. (2002). The effects of mental representation on performance in a navigation task. Memory & cognition, 30(8), 1189-1203.
Farris, Candace, Trofimovich, Pavel, Segalowitz, Norman, & Gatbonton, Elizabeth. (2008). Air traffic communication in a second language: Implications of cognitive factors for training and assessment. Tesol Quarterly, 42(3), 397-410.
Godfrey, J. J., and Linguistic Data Consortium. (1994). Air traffic control corpus complete. Philadelphia: Linguistic Data Consortium.
Gorman, Kyle, Howell, Jonathan, & Wagner, Michael. (2011). Prosodylab-aligner: A tool for forced alignment of laboratory speech. Canadian Acoustics, 39(3), 192-193.
Guion, Susan G. (2005). Knowledge of English word stress patterns in early and late Korean English bilinguals. Studies in Second Language Acquisition, 27(04), 503-533.
Kim, H., & Elder, C. (2011). Understanding aviation English as a lingua franca: Perceptions of Korean aviation personnel. Australian Review of Applied Linguistics, 32(3).
Mampe, Birgit, Friederici, Angela D, Christophe, Anne, & Wermke, Kathleen. (2009). Newborns' cry melody is shaped by their native language. Current biology, 19(23), 1994-1997.
Nazzi, Thierry, Bertoncini, Josiane, & Mehler, Jacques. (1998). Language discrimination by newborns: toward an understanding of the role of rhythm. Journal of Experimental Psychology: Human perception and performance, 24(3), 756.
Ramus, Franck, Nespor, Marina, & Mehler, Jacques. (2000). Correlates of linguistic rhythm in the speech signal. Cognition, 75(1), AD3-AD30.
Tajima, Keiichi, Port, Robert, & Dalby, Jonathan. (1997). Effects of temporal correction on intelligibility of foreign-accented English. Journal of phonetics, 25(1), 1-24.