Deck 6: Speech Perception

Question 1

Which kind of sound contrast differs along the sound dimension that is called "manner of articulation"?

A) Whether air is closed off between the two lips or near the glottis
B) Whether flow is completely stopped or is only partially constricted by a narrowed passage somewhere in the vocal apparatus
C) Whether or not the vocal folds vibrate
D) A contrast caused by the relative height of the tongue during sound production

Answer 1

B

Question 2

Which sound contrasts differ in their place of articulation?

A) /b/ vs. /p/
B) The /p/ in pool vs. in spool
C) oo vs. ow
D) /p/ vs. /k/

Answer 2

D

Question 3

Which response is not a challenge engineers faced in designing automated speech recognition software?

A) The stream of speech sounds proceeds without many silences, making it hard to decide when the signal has ended one word and begun the next word.
B) Individual speakers vary widely in how they pronounce phonemes.
C) Even a voice-activated toy dog must understand the meaning of individual words if it is to respond differently to each of them.
D) Background noise overlaps with the kinds of sounds present in the speech signal, making it unclear which source produced which component of the sound stream.

Answer 3

C

Question 4

Which of the following adds nothing to variability between different instances of the same phoneme in speech?

A) Differences in the vocal apparatus between men, women, and children speakers
B) Morphemic content
C) Coarticulation
D) The allophones included in the spoken language

Question 5

Two or more similar sounds that are variants of the same phoneme are called

A) allophones
B) morphemes
C) warped perceptions
D) sorted phonemes

Question 6

Which speech-perception task would place the greatest demand on short-term memory (working memory)?

A) A task in which a subject in an fMRI scanner passively listens to repeating syllable stimuli that sometimes change acoustically
B) A task in which a subject is instructed to click on a computer image that matches a word they hear
C) A forced-choice identification task
D) An ABX discrimination task

Question 7

In a test for categorical perception, participants hear two sounds and are asked to say whether the sounds are the same phoneme or different phonemes. What have such tests found?

A) Performance is poor at first, but listeners become more discerning between sounds with practice.
B) Performance is poor despite large acoustic differences, as long as the sounds fall within the same phoneme category.
C) Performance is better if the sounds are acoustically different and fall within the same phoneme category than it is if they come from different phoneme categories.
D) Listeners can respond equally well to all sound differences but take more time to respond if the sounds come from the same phoneme category.

Question 8

Suppose researchers ask participants to identify syllables from a set of stimuli that were created such that they vary systematically by equal steps of difference in voice onset time. How well will the participants identify stimuli?

A) Syllables are identified correctly only if they occur frequently in the language.
B) Subjects will be unable to perceive syllables near the VOT boundary between phonemes.
C) On either side of the phoneme boundary in VOT, subjects will consistently hear each syllable as a good example of one or the other phoneme.
D) On either side of the phoneme boundary in VOT, subjects will guess the syllable at close to a chance level.

Question 9

How is categorical perception different from continuous perception?

A) Some exemplars of a stimulus category are perceived as much better than others.
B) Subjects respond equally well to any kind of change along a physical gradient.
C) The proportion of identifications as one vs. another category of stimuli has a very steep slope around some boundary that is not based on physical differences.
D) It involves a continuously updating perception.

Question 10

Which statement accurately describes a similarity or difference in the way acoustic speech and visual sign languages are perceived?

A) Vocally produced sounds vary in several acoustic dimensions, whereas only the shape of a hand gesture matters in sign language.
B) Experienced listeners of spoken language can ignore sound variations, whereas even those fluent in sign language never learn to ignore variation in hand shape.
C) Speakers acquire the accent of their communities, whereas signers produce universally stereotyped hand movements and shapes no matter where they use sign language.
D) There are no major differences-visual perception shares many principles with auditory perception, and both types of language function in an equally complex manner.

Question 11

______ is a clear example of how visual perception can influence how we hear speech.

A) The McGurk effect
B) Coarticulation
C) The Ganong effect
D) Complementary distribution

Question 12

Imagine that you were a participant in William Ganong's 1980 study, and you were asked to report whether you heard the word bid or pid when presented with an ambiguous phoneme sound between /b/ and /p/ at the start of _id. You would

A) be equally as likely to report hearing pid as bid.
B) be more likely to report hearing pid than bid.
C) be more likely to report hearing bid than pid.
D) likely be unable to recognize which word you heard.

Question 13

The Ganong effect is to linguistic perceptual invariance as _______ is to visual perceptual invariance.

A) the uniqueness point
B) implicit priming
C) shape constancy
D) color constancy

Question 14

_______ occurs when a participant reports hearing both the missing phoneme as well as the non-linguistic sound, such as a sneeze, that replaced it.

A) The phoneme restoration effect
B) Compensation for coarticulation
C) A mondegreen
D) Phonemic awareness

Question 15

What determines which sound features become prioritized as cues about sound category distinctions?

A) Cues are weighted differently depending on the context, variability, talker, and language.
B) The most salient features stay the same because the auditory system constrains perception.
C) All cues are on equal footing as important dimensions of speech sounds.
D) Voice-onset time is always prioritized over pitch because pitch has the additional functions of cuing speaker identity or prosody.

Question 16

When testing the perceptual effects of music training, why should a researcher use random assignment of participants to a musical group or control activity?

A) Participants are usually equally glad to be assigned to any kind of activity.
B) Otherwise, the direction of causation is questionable, since those in the music group may already have superior listening skills that made them more likely to want to play music.
C) Random assignment makes the music group more diverse.
D) Random assignment minimizes individual differences within each group, making the groups easier to reason about.

Question 17

Emily Myers (2009) performed fMRI brain imaging on subjects who were hearing and watching stimuli that produced the McGurk illusion. What aspects of the stimuli predicted selective brain activity?

A) Neural responses during the McGurk effect are no different than during trials without the effect.
B) Auditory neural regions represented the phonemic category rather than the acoustic gradient within it.
C) Auditory neural regions represented any kind of perceptible acoustic change.
D) Some neural regions responded selectively to an abstract sound category, whereas other neural regions responded selectively to acoustic details even within a category.

Question 18

When a 50-year-old man says the name Rex, he pronounces the vowel with a lot of intensity at 500 Hz. When a 7-year-old girl says the name Rex, she pronounces the vowel without a strong component in the frequency range around 500 Hz. You have no trouble hearing both of their speech as examples of the same word. This phenomenon is known as

A) perceptual invariance.
B) simultaneous activation of multiple word meanings.
C) incremental language processing.
D) crossmodal priming.

Question 19

While researching a language that linguists have not catalogued before, you discover that there are very strong accents within it. In a paper presenting a hypothesis about the origin of the accents, you use Bill Labov's research in the United States to corroborate your hypothesis. Your hypothesis is most likely:

A) The inherited physical traits of each group's vocal tracts lead them to produce slightly different sounds, giving them an accent.
B) When people move into a new area, they develop an accent that propagates better in that environment, based on its acoustic features.
C) Identifying with an esteemed, yet otherwise threatened, subset of the population can motivate speakers to signal that identity by an accent.
D) Subsets of a population that value education less will acquire an accent out of sloppiness.

Question 20

When dubbing a film into another language has been done very well, the sound team has taken advantage of which aspect of speech perception?

A) You cannot tell, just by looking, how rounded or spread a speaker's lips are held unless you have extensive familiarity with the speaker's face and lips.
B) The McGurk effect works only for vowels, so the soundtrack has more influence than the video image.
C) Viewers take pride in their own language, so they tend to "see" it in all actor speech, especially when the actor is attractive.
D) Voicing and tongue gestures are barely visible, or are invisible, meaning that the actor's appearance is compatible with multiple phonemes.

Question 21

Tasha is a spy who must interact regularly with people from many different parts of the world. Most of her contacts speak English, but they often have heavy accents that make communication difficult. What strategy should she use to become better at comprehending strangers with unfamiliar accents?

A) She should go to subtitled foreign movies to absorb sounds without necessarily trying to comprehend them.
B) She should seek accent-specific training, exposing herself to speakers of each new accents she might encounter.
C) She should expose herself to a number of talkers who each have unique accent, thereby improving her skills at understanding all accents.
D) She must try to obtain recordings of the actual individuals she hopes to be able to understand.

Question 22

Judging from the psycholinguistic evidence, what is a benefit of raising a toddler with exposure to a variety of voices and vocal accents?

A) The toddler will become a better speller.
B) The toddler will distinguish words from nonsense words at a younger age.
C) The toddler will grow up to be speak multiple languages.
D) The toddler will be better at recognizing words spoken with many emotional tones.

Question 23

What general cognitive skill may support adaptation to variations of the sound categories in speech?

A) Managing attention to adjust the weights of multiple sound cues in speech
B) Managing attention to think of how the sound really ought to be produced
C) Refreshing working memory more often by forgetting recent exposures
D) Actively perceiving every acoustic detail as if it might become relevant in identifying phonemes

Question 24

Suppose an individual speaker consistently used VOT to distinguish between two phoneme categories (in her production of voiced vs. voiceless sounds) but was not consistent in cuing by means of the pitch at the beginning of the vowel that follows these sounds. What acoustic cue(s) will this individual then use to categorize voicing distinctions in speech sounds that she passively hears others produce?

A) Pitch
B) VOT
C) Both VOT and pitch
D) Some other cue

Question 25

What was the result when researchers used fMRI to image brain activity while subjects listened to syllables that differed in place of articulation?

A) The motor region for lip movements was more active when hearing /p/, and the tongue movement area was more active when hearing /t/.
B) The motor region for tongue movements was more active when hearing /p/, and the lip movement area was more active when hearing /t/.
C) Brain activity during passive listening did not overlap with the brain regions active during speaking tasks.
D) Upon hearing any speech sound, all the articulatory regions were activated at once, as predicted by the motor theory of speech perception.

Question 26

When is it most likely that articulatory knowledge of how speech sounds are made strongly helps speech perception?

A) When Spanish students learning Basque as a foreign language start by trying to imitate sound contrasts that do not occur in Spanish
B) When people are tasked with manipulating abstract categories of sound, requiring conscious effort
C) When the speech sounds are quite familiar, so the participants know more about their production
D) When the environment is quiet, so listeners can hear the more subtle sounds of jaw cartilage and mouth muscle movements

Question 27

Why might a researcher be justified in wanting to use transcranial magnetic stimulation (TMS)?

A) If correctly targeted, TMS can generate auditory hallucinations of whatever speech sounds the researcher is interested in investigating.
B) When "naturally occurring" damage interferes with brain activity, it often affects many brain regions beyond those the researcher can hypothesize about.
C) The pulses of magnetic stimulation are totally reversible and will not affect the functioning of the brain after they are turned off.
D) Brain disorders do not occur naturally in children during the ages when they are learning to produce speech.

Question 28

Why is it generally harder for individuals older than 70 to understand audible speech?

A) Hearing loss, which is nearly universal in this age group, explains the major difficulties.
B) There are so many redundant cues in speech that it confuses older listeners.
C) Since their own muscles are slowed down, they stop understanding visual cues to lip-read.
D) The processing of temporal auditory information can be impaired in older adults even when frequencies are reliably heard at low volumes.

Question 29

What would be the best stage name for a ventriloquist, in order to best support the illusion that the dummy is the one saying the name during the script?

A) Bob, because the ventriloquist could simply substitute an altered /d/ for the /b/.
B) Pip William, because it is long and with categorically perceived consonants.
C) Henrietta, because it is long and the lips do not have to move.
D) Famous Mave, because this name takes full advantage of the Ganong effect.

Question 30

In general, ventriloquists seem to use techniques and illusions that heavily rely on their audience's _______ processing of stimuli, where previous knowledge has a great influence on shaping perception.

A) top-down
B) bottom-up
C) architectural
D) crossmodal

Question 31

Suppose you are trying to automate the production of speech using computer software and a loudspeaker. Why would it fail to sound natural if your message was generated by splicing together exemplar phonemes that had been spoken one at a time in isolation on a recording?

Question 32

Refer to the figure.
If the figure refers to the categorical perception of /ba/ and /pa/, summarize how a person would interpret these sounds as the VOT changed for each.

Question 33

Refer to the figure.
Based on the figure from Kuhl and Miller's 1975 study of categorical perception in chinchillas, what can you conclude about humans' and chinchillas' abilities to detect sound boundaries and language?

Question 34

Describe examples of evidence that show how context can help a listener comprehend specific phonemes or syllables.

Question 35

Describe the McGurk effect and how it works.

Question 36

Under what conditions does the Ganong effect disappear? What does this tell us about the effects of context on word perception?

Question 37

You move to New York City and have trouble understanding a salesclerk in a local store who has a thick foreign accent. Several months later, your trouble has mostly vanished. Explain what this reveals about peoples' sensitivity to acoustic details in speech.

Question 38

Refer to the figure.
Figure. Average looking time was measured while broadcast phonemes either alternated (between Hindi /d_/ and /dj/) or stayed the same (repeating /d_/, or repeating /dj/). Observations were taken when infants sucked on either a U-shaped teether that allowed free tongue movement (left bar) or a flat teether that immobilized subjects' tongues (right bar). Scores greater than zero indicate a preference for the alternating phonemes. Error bars denote Standard Error of the Mean. The asterisk marks a significant difference from zero.
Infants spend more time looking at a loudspeaker, on average, if they detect an acoustic change within a stream of repeating sounds that it broadcasts. Based on the figure from Bruderer et al.'s 2015 study of speech perception by 6-month-old English-speaking infants (above), does the articulatory system of infants influence whether the babies hear phoneme distinctions between non-English consonants? Explain briefly.

Question 39

How does a ventriloquist exploit features of speech perception to create the illusion that the dummy is the source of the speech the audience hears? Be specific in referring to several of both the perceptual phenomena involved, and the on-stage strategies at play.

Question 40

There is evidence that developmental dyslexia is related to deficits in speech perception. Name specific examples of these deficits that could lead to dyslexia.