Captain Psychology

by

Speech Perception

Speech Perception

  • 3 components of Speech Perception

    • 1. Physical signal

      • Process sound waves

      • Varies in 3 parameters

        • Amplitude

        • Frequency

        • Time

    • 2. Extract  phonemes

      • The smallest unit in a language that is capable of conveying a distinction in meaning

      • Make fine distinctions between similar patterns of sound

      • Examples

        • Buh or Tuh

        • M of mat and B of bat

      • Phone = a particular sound used by any language

      • eg the sound [r]

      • Phoneme = a sound used in contrast to another in a particular language

        • eg  the category  /r/ as distinct from /l/ 

        • 1. Phoneme extraction is categorical

          • i.e. if physical characteristics of the signal are changed slowly, there is a sudden change in which phoneme is perceived

        • 2. The speech recognition system can modify fuzzy input to give the listener the correct sound

    • 3. Perception

      • Extract meaning

      • Allows us to recognize the same sounds spoken in different ways

        • e.g. by two different people

      • There are no natural breaks in speech

      • We “hallucinate” word boundaries

        • Oronyms

          • Two or more sentences that use the same sounds but have different words

          • Such as

            • Scuse me while I kiss the sky

            • Scuse me while I kiss this guy

          • Or

            • Some others I know

            • Some mothers I know

        • Speech perception errors

          • Eugene O’neil won a Pullet Surprise

Research methods

  • Voice-Onset Time

    • The amount of time between the release of a stop consonant and the onset of glottal vibrations in the following vowel

    • In English, if we start the laryngeal tone exactly at the beginning of the “P” sound, it becomes (is perceived as) the “B” sound

    • If we delay progressively longer in small increments the beginning of the voice (voice onset time), there is a point in time that it would become the “P” sound

    • VOT may be negative, zero or positive

      • Zero

        • Vocal-cord vibration has begun simultaneously with the release of the plosive consonant

        • A voiceless unaspirated stop (eg. [k]) has zero VOT

      • Negative

        • Vibration beginning earlier than the release

        • A pre-voiced stop (eg. [ɡ]) has negative VOT

      • Positive

        • Vibration beginning after the release

        • A voiceless aspirated stop (eg. [k ʰ]) has positive VOT

    • Different languages have different methods of phonetic realization of this feature

Categorical Perception

  • Definition

    • Sharp phoneme boundary

    • Discrimination peak at phoneme boundary

    • Discrimination predicted from identification

      • only “different” if different phoneme

  • Occurs with consonants, not vowels

  • Not restricted to speech

    • Also found in comparison of musical intervals

  • Not restricted to humans

    • Chinchillas and quails show the same Voice Onset Time boundary as humans

    • Macaques show discrimination peaks at human VOT and place-of-articulation boundaries

  • Innate & Acquired

    • Infants born with ability to make many speech discriminations that they can subsequently NOT make

    • Adults have lost the ability to make distinctions that their language does not use

  • Each language has its own distinctive set of phonemic categories

    • English distinguishes /r/ from /l/  but Japanese doesn’t

    • Tamil distinguishes dental /t1/ from an alveolar /t2/ from a retroflex /t3/.  English doesn’t.

  • Phonemes in a particular language are defined by minimal pairs

    • i.e. since in English “lice” and “rice” have a different meaning, then they contain different phonemes: /l/ and /r/

    • But there is no such minimal pair in Japanese, so they have a single phoneme /r/

  • Can Japanese really not hear any difference?

    • For English speakers /d/-/g/ boundary is in a different position after /l/ than after /r/.

    • This is also true for Japanese who can hear /r/ vs /l/

    • But ALSO true for those who can’t.

    • Is this because of language knowledge (implicit phonetics)? No, its phoneme distinction.

    • QUAILS DO IT TOO  !!!

Auditory Agnosia

  • Definition

    • The defective recognition of auditory stimuli in the context of preserved hearing – as tested with audiometry

    • Primary signs

    • Difficulty in understanding the meaning of spoken words

    • Can refer to a generalized disorder affecting perception of all types of auditory stimuli including non-verbal sounds, speech and music

  • Associated with bilateral, or unilateral lesions of the left superior temporal cortex

    • Although some cases have been described following unilateral right temporal lobe damage

  • By far the most common cause is cerebro-vascular accident

    • But some cases have been reported following encephalitis

  • Types of sound recognition disorders

    • 1. Apperceptive

      • Impaired acoustical analysis of the perceptual structure of an auditory stimulus

      • (frequency, pitch, timbre)

    • 2. Associative

      • An inability to associate a successfully perceived auditory stimulus with a conceptual (semantic) meaning

  • Spoken Word recognition

    • Morton’s 3-stage model

      • Auditory analysis system = identifies phonemes in the speech wave

      • Auditory input lexicon = identifies the phonological properties of known words

      • Semantic system = identifies the meanings of known words

‘There are two great principles of psychology: people have a tremendous capacity to change, and we usually don’t.”   Ken Tangen