After encountering a few students, both native and non-native speakers of English, who found it very difficult to label syllables as stressed (primarily or secondarily) as opposed to unstressed, Laura Koenig (firstname.lastname@example.org) posted a request on the LINGUIST listserv for hints on how to make students aware of lexical stress patterns in English. .She received about 40 replies ranging in content from clever classroom tricks, to lists of useful examples, to thoughts on how the linguistic and phonetic characteristics of stress might be relevant considerations.
Laura states, "The quantity of information also seemed to call for some degree of summarization and paraphrasing; I hope I have not thereby misrepresented anyone's opinions. This has been a useful and interesting activity for me, and I hope the following summary may be of some interest and/or practical use to others."
Her summary, organized into 4 sections follows:
1. Hints in 3 categories
2. Materials: Texts, minimal pairs, word lists, Buddy Holly.
3. Further commentary on the intuitive obviousness of lexical stress patterns
The hints for pointing out stressed syllables to speakers basically fell into 3 categories:
1) Those that used linguistic materials;
2) those that in some way invoked musical techniques;
3) those that involved performing some motor act in conjunction with the word.
In the first category, Bruce Nevin most concisely sums up the spirit of tapping linguistic intuitions: "Be Socratic, ask them what's the difference between present (n and v)." Sean Jensen offered similar thoughts as a 'phonologist's approach.' Many respondents felt that minimal pairs in a known language were the best way to focus attention on stress and/or engage student intuitions. A list of the pairs they suggested is provided in the materials section below.
Frank Blair noted that the English noun-verb pairs are a useful way of showing students that they do, in fact, react to stress patterns in everyday linguistic situations, even if they have difficulty identifying stress patterns otherwise. Don Churma commented, however, that some speakers may be losing the noun-verb stress contrast.
Margaret Fleck suggested that putting words in sentential context may make students more aware of errors in stress; thus sentences like I started typing on the con 'sole (v) may sound particularly odd. Tom Cravens uses minimal pairs to generate class discussion: "Write _invalid_ on the board, and ask them how it's pronounced..." Another possibility is presenting them with unknown words and asking the pronunciation: "[Tell] them that they have to announce the home towns of distinguished Italian visitors at a banquet... [or] prepare cards for someone to use in doing such announcements, marking the stressed vowel." See his list of place names below under materials.
Allan Wechsler recommended that "foreign students whose native language has contrastive stress (say, Spanish) should work with examples in their own languages, to better engage their intuitions."
Ralf Grosserhode & Frank Blair suggested using nonsense polysyllables varying in stress placement in conjunction with the real language materials. Somewhat interestingly (?), they both supplied, specifically, trisyllabic CVCVCV's...
Peter Binkert reported success in having students practice reading aloud sets of words that vary systematically in stress pattern. For example, students begin by reading words with the pattern WSW, then go on to SWW words, etc. He kindly included an extensive listing with his response, which is reproduced in full [less one word that was unrecoverably warped in transmission] in the materials section below.
Don Churma mentioned that derivationally-related pairs that show stress shifts might also be useful illustrations (e.g., celebrate - celebration).
Sharon Flank passed on a trick used by a Polish instructor she knew: "Have students repeat only the vowel pattern of words, with the correct stress pattern, e.g. Massachusetts ae - uh - OO - E
This seemed to work even for those poor souls who were otherwise unable to repeat the polysyllabic Polish words - in part, I suppose, because they didn't need to concentrate on nasty consonant clusters." [It occurs to me that this might be a nice way in English of highlighting the relationship between stress and vowel quality.]
Finally, a few persons suggested that poetic language might help focus students' attention on stress patterns.
James Kirchner: Find a short poem that won't track properly if some of its words are replaced with others
Marie Egan: "Read your class some Dr. Seuss, Shel Silversten, or something like that." Similarly, from Mark P. Line: "use some of the Dr. Seuss rhymes that over-emphasize rhythm and stress. You know the ones..."
A closely related tactic uses on chants and musical lyrics, as the following (again supplied by Mark Line):
YAN kee DOO dle WENT to TOWN
a RI ding ON a PO ny
he STUCK a FEA ther IN his CAP
and CALLED it MAC ca RO ni.
Get them to try to sing it with the wrong stress. Can't be done.
He pointed out that this approach essentially means working from phrase- level stress down to word-level stress. Deborah D. Kela Ruuskanen also suggested introducing the topic of stress using phrasal examples such as YOU're not going out. versus You're NOT going out!, etc. Finally, I include in this context the sentiments of Jefweb@aol.com: "Perhaps, like music, teaching stress may be more successful at times by teaching the phrase pattern as an important unit -- like learning music note by note verses motivating the phrase."
Bret Parker cited TESL works by Carolyn Graham as a possible source of chanted materials (see below under materials) and wondered if a friendly cadet might be able to supply some polite marching chants. Peter Ladefoged also mentioned marching: "Usually [people] will make the stressed syllable coincide with a step--if they are speakers of English."
A few folks went the whole 9 yards [where does this phrase come from, anyway?] and suggested drawing very explicit connections between music and speech. Marion Kee surmised that the problem of "hearing stress" might be similar to the problem a few people have of "picking up a rhythm," and wondered whether music teachers might have some useful suggestions. Bret Parker pointed out that "word-stress and musical settings are fairly consistent... [I]n a hymnal with over 300 hymn settings, not a single hymn ... has an initial word "a," "an," or "the" starts on the downbeat (the beat of principal stress). So I like to teach people about how stress falls in a measure, that in common time (4/4), there is a principal stress on one and a secondary stress on three. Beats two and four get minimal stress."
Similarly, Fred Field said, "I'll start with an even, alternating rhythmic pattern with my hands (right-left, right-left, and so on). To help children in their first music lessons, teachers will accent the first beat. So, you have a pattern like X-x-x-x, X-x-x-x, something like that. By shifting the stress around (for example, to the second beat), you create a different rhythm. I then explain that language (i.e., English) is similar. We establish a rhythm in our speech. Most students have heard rap with very rhythmic speech... "
[In retrospect, I'm a bit surprised that this was the only mention of rap I received. It's certainly a relevant example.]
Fred also passed on an interesting anecdote about variation in audiences clapping to Buddy Holly, which is included at the end of the materials section. [I'm not sure whether the example tells us something about cultural differences in musical form, or something about dialectal differences in linguistic rhythm, or what...but anyway, take a look at it.]
The other major musical technique involved relying on pure pitch or, more rarely, intensity changes to indicate stress. Four persons (Kevin Russell, ASP Elissa (EASP@shark.stmarys.ca), Karen Stanley and Susan Ervin-Tripp) recommended having students hum words rather than speak them. This may help demonstrate multiple stress levels: The highest pitch indicates primary stress, a mid tone indicates secondary stress, and low pitches occur on unstressed syllables. (In this context, I should mention that Noriko Watanabe (email@example.com) requested a summary of the replies, in the hopes that some hints on teaching stress might assist her in teaching native speakers of Japanese to locate where the pitch accent occurs...) Deborah D. Kela Ruuskanen recounted a professor playing musical clichˇs on a violin (e.g., the opening sequence to Beethoven's 5th: da da da DAAA) to Chinese students to illustrate the notion that stressed syllables in English varied in length and pitch. Finally, Doug Honorof and Richard Cameron suggested [I've got to admit, I can't wait to try this one] speaking words through a kazoo, to obliterate segmental information, emphasize loudness differences between stressed and unstressed syllables, and, possibly, force the speaker to exaggerate pitch contours.
My original query on LINGUIST mentioned a method that I had learned some years ago: Tap once as you say a word; most speakers will tap on the stressed syllable. Peter Ladefoged offered an explanation for this phenomenon: "...if you take stress to be something that a speaker does (rather than a phonological construct), it is only on these syllables that people can recognize that they are doing something (putting in extra respiratory effort), and only at these moments can they synchronize another bodily movement, such as a tap." At the same time, Marion Kee cautions that some people may require much practice before they are able to synchronize body movements and speech. Several respondents mentioned having learned the tapping method; Erika Konrad mentioned it in connection with Gilbert's ESL textbook (see materials section), and James Kirchner reported having learned it in 1st grade. Erika Mitchell reports having learned as a child "[to put] our fingers on our throats and feeling (or trying to feel) for the extra tension and release involved in the stressed syllable. It was a far from perfect method, but with practice, we could answer our little stress worksheets with better than chance accuracy." [There is a clear parallel here to the old method of discerning tense and lax vowels--but that's its own debate.]
There were some interesting variations on the tapping theme as well:
a) Jean-Louis Duchet uses the following with French students of English and phonetics: Starting with an actual word, tap out its rhythm, with one tap for every stressed syllable and a stronger tap for the stressed one. Then have students reproduce the tapped rhythm themselves. Finally, have the students say the word as they tap.
b) Haj Ross passed on his "tapping with a wrinkle" method: "Get the students to say the word fast 5 - 10 times, hitting the table with their hands on each repetition. The faster they go, the more the stressed syllable will be what they hit on. Then make them slow down the pronunciation, while still repeating, until they are going one syllable / second, when they either hear the accent or fail, right?"
c) Kate McCreight dispenses altogether with the subtleties of small motor acts: "Have them pretend they are actors in a bad movie, and the word in question is some kind of nasty name: "why, you dirty ________!" Have them pound the table or punch the air as they say it. Pounding seems to work better than tapping."
Finally, Jessica Williams, Doug Honorof and Richard Cameron described a technique sometimes used in ESL classrooms--a coincident act that may serve as visual illustration as well: Use a rubber band (either in your hands or the student's) to demonstrate stress. The usual tendency is to stretch the rubber band on the stressed syllables, and retract it on the unstressed ones. Richard noted that this method demands a fairly slow speaking rate.
A. Books, texts, etc:
Bret Parker provided the following reference:
Graham, Carolyn. _Jazz chants for children: rhythms of American English Through Chants, Songs, and Poems_. New York: Oxford Univ Press, 1979.
and surmised that this author may have done works for adult audiences as well.
The most widely cited ESL text (Erika Konrad, Karen Steffen Chung, and Richard Cameron) was Judy Gilbert. _Clear Speech_. Cambridge University Press which teaches stress by the tapping method and also by definition (length, pitch and loudness compared to surrounding syllables).
Another suggestion from Karen S. Chung was Linda Grant. _Well said_. Heinle and Heinle publishers.
Don Churma recommended--despite a possible counter-example--the Akmajian et al. Intro textbook for a Liberman/Prince account of stress shifts through derivation in English (e.g., celebrate-celebration).
B. Word pairs:
German (thanks to Ralf Grosserhode)
Umfahren - to drive over s.o.
umfAhren - to drive round s.th
Dutch (thanks to Dirk Janssen)
bed_e_len - to give everyone his/her part
b_e_delen - to beg
Italian (thanks to Tom Cravens)
par 'lo - he/she spoke
'par lo - I speak
'prin cipi - princes
prin 'cipi - principles
English pairs (thanks to Margaret Fleck, Dirk Janssen, James Kirchner, Frank Blair, Tom Cravens, Peter Binkert, Bruce Nevin, Don Churma, Rosa Graciela Montes, Deborah Bobrow). [The following "pairs" vary greatly in the degree of reduction of the unstressed vowel. One might take the position that true 'minimal pairs differing in stress' show little vowel quality variation, e.g., permit, survey, pervert (n-v). Alternatively, of course, one could say that vowel quality/reduction is simply one component of English stress.]
noun-verb pairs (Noun pattern = SW, verb pattern = WS)
Is it possible to __________ this file?
You better put a __________on that until you get to the doctor.
Alice's major is _____________.
I'll have to __________ the car.
The Gettysburg ___________ What's your __________?
A rare pair:
A pair that may exist for some speakers:
admit (Deborah Bobrow has seen this form used as a noun used in writing to mean 'admission.')
noun-adjective pairs :
invalid (n vs. adjective)
Since the accident, Elmer has been an ______________.
That parking sticker is __________; it says "good until August 31, 1994!"
'con tent , con 'tent
'per fect , per 'fect
specialty, special tea
Phrasal pairs varying in stress pattern (to differentiate primary and secondary stress)
White House vs. white house
the 2 hotdogs vs. the 2 hot dogs
[Everybody CAN tell the difference...or they might be served a canine with an elevated temperature, rather than a "HOT dog"! -D. Churma]
Cases where stress may vary among speakers, or fall in an unexpected location:
De 'troit vs. 'DE troit
ci 'gar vs. 'CI gar
New Berlin IL (new 'berlin)
Bolivar MO ('baliver)
Piasa IL ('paiesa) (e=schwa here)
Benld IL (b'neld).
C. Tom Cravens's list of Italian place names illustrating stress patterns:
D. Words grouped by stress pattern (thanks to Peter firstname.lastname@example.org) [In looking this over I am reminded of rhyming dictionaries...]
VOWELS, STRESS, AND SPELLING: Transcribe the following words including placement of primary stress. (v) =verb; (n) =noun; (a) =adjective
(1) Two syllables:
agree, debris, ennui, Pawnee, marquee, esprit, asea angry, money, sunny, gypsy, litchi, Fifi, squeegee survey (v), Bombay, bouquet, inveigh, parfait, souffle', soiree survey (n), bomb bay, Tuesday, foray, subway, cadre, entree' review, canoe, shampoo, construe, eschew, ragout, undo preview, igloo, curfew, yahoo, Zulu, voodoo, cashew below, outgrow, plateau, Bordeaux, although, jabot, hello bellow, meadow, aloe, hobo, Margot, cocoa, de'pot, outcry (v), apply, goodbye, sky high, descry, untie, Versailles outcry (n), ally, cacti, Ely, magpie, fish fry, pigsty soda, tchotchke, bwana, chutzpah, geisha, djibbah, schmatte
(2) Three syllables:
broccoli, apogee, fricassee, apathy, symphony, pedigree, ivory shillelagh, chianti, bologna, Delancey, obliquely, confetti, Hatari guarantee, bourgeoisie, chimpanzee, Tennessee, fleur-de-lis, potpourri, jamboree, runaway, Saturday, ricochet, sobriquet, protˇgˇ, workaday matinee, San Jose, Beaujolais, Chevrolet, Santa Fe, negligee, overplay, rendezvous, parvenu, residue, peekaboo, interview (n), honeydew, avenue Timbuktu, kangaroo, misconstrue, overthrew, hitherto, buckaroo, overdo mistletoe, Mexico, gigolo, furbelow, embryo, cameo, Scorpio fiasco, concerto, rococo, soprano, falsetto, crescendo, alfresco portmanteau, overflow (v), apropos, status quo, tallyho, cheerio, to-and-fro amplify, rockaby, occupy, incubi, prophesy, alkali, underlie subpoena, agenda, charisma, pariah, pagoda, kimono, babushka
(3) Four syllables:
apostrophe, Menomini, Penelope, fraternity, hyperbole, anemone, telegraphy Ypsilanti, necromancy, kamikaze, Cincinnati, insincerely, poison ivy, hokey-pokey pistachio, Arapaho, imbroglio, Lothario, simpatico, adagio, portfolio, armadillo, Filipino, San Marino, virtuoso, Sacramento, lucky fellow, punchinello Arabia, azalea, Bulgaria, agraphia, et cetera, millennia, America, hullabaloo, Kalamazoo, Tippecanoe
E. The Buddy Holly anecdote (from Fred email@example.com)
In the early days of rock n' roll, so-called Black audiences listened to Black artists, and white folks listened to their own type of music. In those days Black music was called rhythm and blues. When artists like Holly came on the scene--and later Elvis--fans were somewhat confused because they couldn't tell if these artists were black or white. Holly was asked to play at the Apollo--a famous Black nightclub in Harlem--before it was discovered he wasn't black. He was later invited to play, as the story goes, in Atlanta. Black radio stations played his music as did the white stations. At his concert, both groups showed up, and in the atmosphere of segregation created a large mess. Organizers needed to either tell one group or the other to go home, or arrange some type of seating that would maintain the outward appearance of segregation. So, Black folks sat on one side of the auditorium, and White folks on the other. Cultural patterns showed up right away, with the African-Americans clapping to the music on 2 and 4 (the second and fourth beat in the musical measure), and the Caucasians clapping on 1 and 3. The clapping alternated from one side of the room to the other.). The moral to the story is that we clap on the stressed (or accented) beat in the music, which when accompanied by lyrics means the stressed syllable.
3. Further commentary on stress and its intuitive obviousness
As for the overall issue of students finding it difficult to perceive stress contrasts, my respondents represented the full spectrum from incredulity to empathy. The notion that non-native speakers might have difficulty was generally accepted, and it even produced a classroom technique: "Play tapes of the speech of two native speakers, one of whom, however, misplaces stress on a couple of items. Tell them that one is a very fluent non-native speaker, and ask which one" (Tom Cravens). But the native English speakers were another matter entirely; cf. James Kirchner's comment, "I'm pretty amazed that any native-speaking undergrads can't hear English lexical stress, but I guess anything's possible." Allan Wechsler suggested that non-native speakers might have difficulty if their native language did not use contrastive stress:
"Listen to their own pronunciation of English. Do they _produce_ stress correctly? If not, then they have simply never internalized the notion. If they are to learn it, they must do so intellectually, the way we learn all foreign phonology." But again, native speakers were another story: "The native English speakers -- wow, that's a toughie. Have you tested their ability to _count_ syllables? Can they say words syllable-by-syllable? I'm tempted to suspect a real neurological deficit, and I'd look for other clues."
Many responses implied that the goal with native English-speaking students was essentially to make their intuitions obvious to them; cf. Frank Blair's observation that "once folk are convinced that they can actually hear stress, the rest becomes easier." Peter Ladefoged felt that one couldn't expect persons to be able to apply a "stressed" label to syllables other than those with primary stress and the one with the phrasal (pitch) accent, but "even that is very difficult for non-native speakers."
In the meantime, bits like "I finally found something that works" (Haj Ross) and "Tough subject!" (Kate McCreight) were helping convince me that I wasn't creating mythical difficulties in these few native speakers (especially after a discussion one morning over breakfast with Arthur Abramson, who fell squarely among the incredulous). A few persons "confessed" to having their own problems identifying stress. After proposing that rhythmic poetry might help students perceive stress patterns, Marie Egan added that this ability might not transfer to other situations. "When I had to write poetry in college I looked up every multisyllable word in the dictionary to find out its stress." Don Churma outlined a couple cases where subtleties of phonetic vowel quality may make stress identification difficult "even" for phonologists and phoneticians. One of these was the difference between an unstressed vowel and a non-primarily stressed /I/. [I was rather happy to see this one; I have _always_ been skeptical how /I/ and /U/ play into stress-level differences where vowel quality is supposed to be a major cue.] In a similar vein, Mark Line observed the particular difficulties posed by running, as opposed to citation form, speech: "I think a lot of what we perceive as lexical stress as native speakers is often completely unreflected in the signal. We know the word's stress pattern (when uttered alone, for instance), and so that's how we perceive it." [Indeed, there are times when I find myself in quite a conceptual struggle trying to reconcile the notion of linguistic stress levels with the phonetic nature of stress as a gradient and extremely -- shall we say 'plastic' -- phenomenon]. The final word on the theme of variance I give to Janet Pierrehumbert: "...some people simply do not hear stress well; some experiments by Nakatani and Aston suggest that stress is just not very salient to some people. We got a lot of interspeaker variability in affects of stress in our article that just came out in Language and Speech (Pierrehumbert and Nair). It's kind of shocking to have such a central concept exhibit so much interspeaker variability, but there you are."
A few respondents reported similar difficulties teaching stress, or related notions, to native speakers of other languages. Deborah D. Kela Ruuskanen related difficulty teaching stress in Finnish, and Noriko Watanabe has encountered some native Japanese who have a hard time perceiving pitch accent. Clearly there is _a lot_ that can be said (and questioned) about how stress and its perception may vary as a function of language, but that takes us a bit afield of the question at hand. I will, however, pass along the following account from Dan Everett:
"Your query raises a fascinating issue that Peter Ladefoged and I dealt with this summer in our investigations into the stress patterns of Banawa and Piraha. In both cases we were concerned with checking out (my own) published claims on the stress patterns of those languages.
With Banawa, it turned out to be quite easy to teach three men to tap with their fingers, or with a pen on a metal bar, the stressed syllables. They were all quite accurate and consistent.
With Piraha, however, we attempted to teach this to three people. Two of the three couldn't get the point at all and either tapped randomly, *after* saying the word, or on each syllable in careful, slow speech. One person was able to learn to do this but not consistently. Sometimes he would tap after the word and sometimes only at the end of words. But in several cases, mainly when he wasn't tired, he tapped on the stressed syllables.
I suspect that the differential success we had in Banawa and Piraha is related to the fact that Banawa is not tonal, its stress is postlexical, and it has a fairly constant acoustic correlate (loudness) - but I am only speculating (Piraha is tonal, so stress must compete with tone for prominence - since tone is lexical, stress is less salient than tone; in Piraha morphology crucially interacts with stress; in Piraha stress doesn't have a constant acoustic correlate - can be loudness, length, or tone intensification).
We therefore tried another tack with Piraha. We recorded a native speaker pronouncing the words we wanted to test. Then we computationally manipulated the amplitude, length, pitch, etc. of different words, played these to another native speaker, and to the original informant, and asked them to tell us which ones were good and which ones were "crooked" (Piraha expression). This worked much better. But it still wasn't 100% consistent.
So I suspect that it will be easier to teach native speakers to identify stressed syllables when there is "less prosodic competition" in the utterance.
In any case, you are absolutely correct - teaching stress is not as transparent as some parts of the literature suggest.
Thanks to the following (roughly in order of receipt of initial message):
Bret Parker, University of the Pacific libraries
Ralf Grosserhode (Afrikanistik2@uni-bayreuth.de)
Dan Everett (firstname.lastname@example.org)
Peter Binkert (email@example.com)
Margaret Fleck, UofIowa Comp. Sci. (firstname.lastname@example.org)
Dirk Janssen (email@example.com)
Jessica Williams, UofI @ Chicago (U17883%UICVM.bitnet)
Allan Wechsler (Wechsler@world.std.com)
Marie Egan (firstname.lastname@example.org)
Tom Cravens (email@example.com)
Frank Blair (firstname.lastname@example.org)
Erika Konrad, Northern Ariz. U English (email@example.com)
Karen Stanley, Central Piedmont Comm. Coll. (firstname.lastname@example.org)
Noriko Watanabe, U of Oregon East Asian Lang&Lit (email@example.com)
Kevin Russell, U of Manitoba Linguistics (firstname.lastname@example.org)
Peter Ladefoged, UCLA (DU0PNL@mvs.oac.ucla.edu)
James Kirchner (JPKIRCHNER@aol.com)
Fred Field (email@example.com)
Sharon Flank, Systems Research & Applications Corp (firstname.lastname@example.org)
Jean-Louis Duchet, Universite Paris7-Denis Diderot & Universite de Poitiers (email@example.com)
Deborah D. Kela Ruuskanen (firstname.lastname@example.org)
Bruce E. Nevin, Cisco Systems (email@example.com)
ASP Elissa (EASP@shark.stmarys.ca)
Mark P. Line (firstname.lastname@example.org)
Janet Pierrehumbert (email@example.com)
Haj (John Robert) Ross (firstname.lastname@example.org)
Don Churma (email@example.com)
Kate McCreight (firstname.lastname@example.org)
Rosa Graciela Montes (email@example.com)
Karen Steffen Chung, Nat'l Taiwan U. (firstname.lastname@example.org)
Richard Cameron (U17819%UICVM.bitnet@YaleVM.CIS.Yale.Edu)
Deborah Bobrow, Berkeley Systems, Inc. (email@example.com)
Susan Ervin-Tripp (firstname.lastname@example.org)
Marion Kee, Carnegie Mellon (Marion.Kee@a.nl.cs.cmu.edu)
Sean Jensen, School of Oriental & African Studies (email@example.com , firstname.lastname@example.org)
Erika Mitchell (email@example.com)
Thanks also to Doug Honorof for non-electronic discussions on this topic.
laura l. koenig firstname.lastname@example.org Thank you for your wonderful post and summary on stress, Laura. I maintain a homepage on communication disorders. Would you allow me to placethe information in a file, citing you of course, and place it in my home page which is located at http://vax1.mnsu.edu/~kuster/Welcome.html or will you be placing it someplace on the Internet so I can point to it? It certainly has therapeutic relavance, as well as relevance for those of us who teach phonetics, or those involved with accent reduction. Looking forward to hearing from you. Judy Kuster email@example.com Exit