Martello tower, photo by Erik Simpson

Vowel Alternation in Ulysses

Linguistic Background: What Is a Vowel, After All?

A vowel is a sonorous speech segment, occurring at the nucleus (often the middle) of a syllable, "produced without any kind of obstruction of the outgoing breath" (Ladefoged 26). If a speech segment is unvoiced—that is, articulated without vibration of the vocal cords—or involves any obstruction to the vocal tract (at the lips, teeth, palate, or throat) we deem it a consonant rather than a vowel. Whereas the wave form of a stop consonant like [p] or [t] is an abrupt blip, and the wave form of a fricative consonant like [s] or [f] looks like random chaotic squiggles and zig-zags, the wave form of a vowel shows regular periodicity with generally higher amplitudes, indicating its sustained sonority. To get an idea of what this looks like, see the Vowels Appendix for an image of the wave-form from the vowel [æ] as in "quack."

How do we describe the differences between vowels?

There are four main parameters we use to identify vowels; three relate to the shape of the vocal tract (vowel articulation), the other is based on the duration of the sound (vowel length). In terms of length, we can distinguish between short vowels and long vowels (marked with a doubled vowel or a ‘:’ symbol), where the sound is approximately twice as long and the articulatory quality remains constant across the whole vowel. We can also distinguish between vowels with consistent articulation, called monopthongs, and longer dynamic vowels that require an intermediate change in articulation, called dipthongs. Dipthongs are represented as two adjacent vowel symbols, or a vowel plus a glide (‘w’ or ‘y’ sound), where the first symbol indicates the initial articulatory features, and the second symbol indicates the changed quality in the latter half of the sound. In terms of articulation, we can determine a value for lip roundedness (rounded/unrounded), for vertical tongue-position (high/mid/low, or close/mid/open), and for coronal tongue-position (front/central/back). See the Vowels Appendix for a chart of the International Phonetic Alphabet symbols and the articulatory feature combinations they represent. To assess these values empirically, linguists take frequency measurements at the center of the vowel waveform—the F1 (first formant) measurement corresponds to tongue height, and the F2 (second formant) measurement corresponds to a combination of tongue-backness and lip-roundedness.

Although we tend to think about the vowels of a language as discrete contrastive sounds, in reality it is more accurate to define an individual vowel as occupying a particular region of the "vowel space." Linguists represent the space graphically by plotting vowel quality data on a coordinate plane, where the x-axis indicates first formant value and the y-axis indicates second formant value. This strategy for visual organization reveals the boundary between vowel-regions to be flexible and nebulous. So, while we can use featureal and orthographic distinctions to discuss patterns in articulation and quality among sounds perceived as different phonemes (this idea is called the Theory of Categorical Perception), attempting to narrowly or confidently formalize the identity of a specific vowel remains unfeasible. Some of this variance can be attributed to anthropological causes, such as a difference in head size (and thus, vocal tract), or a difference in social/regional identity (divergent dialects).

For many languages, like Spanish, Swahili, and Japanese, it is possible to compile a complete vowel-contrast set, containing (as in a minimal pair) words with identical consonant structure but which representing the full gamut of vowel-sound possibilities. However, according to Ladefoged, this is impossible for English, which has a much larger vowel inventory than the basic 5-vowel a, e, i, o, u system: "we cannot find a single set of English words differing only in the vowels as we did for most of the languages" (28). We should not, therefore, expect to find any exhaustive vowel-alternation sets in Ulysses; but such impossibility/incompleteness (with the possibility for completeness imaginable, but not practically realizable) places this phenomenon among the ranks of typical Joycian antics.

Standard varieties of American and British English distinguish 14-15 and 20 different vowel sounds, respectively (Ladefoged 31). According to Cook (1997), Irish English "pure" vowels (monopthongs) are largely similar to American and British vowels, with perhaps more allophonic variation in the a-like vowels; however, Hickey (2005) suggests that Dublin vowels in particular may not be quite so akin to their British and English counterparts. Of course, such information about precise regional articulation is of limited use toward Ulysses, since language change happens at a rapid pace and these descriptions are based on data from a century later than the speech in the novel. See the Vowels Appendix for more information on Irish/Dublin vowels, as well as a reference catalogue of British and American English vowel inventories.