Five Principles enhancing Spoken Language Comprehension in Healthy Older Adults

(Wingfield, 1999)

Arthur Wingfield wrote an interesting review of the literature on the comprehension of speech in older adults (This was in the context of a text on the use of face-to-face interviews in survey research with older adults).

He first considers the perception of speech (which seems easy and almost effortless to us much of the time):

• The rapid rate at which speech arrives is one important consideration
• While there is no “normal” speech rate, speech rates can range from 90 words per minute in thoughtful conversation to higher than 210 wpm in reading aloud a prepared speech
• Speech perception can be viewed from both “bottom-up” and “top-down” perspective
• Bottom up begins with an acoustic waveform and ends with comprehension of the message at the level of discourse: “You talk, and I’ll listen.”
  • 1. Pholological analysis and segregation of the speech stream
    • speech tends to run together without clear separations between words
    • natural speech is not well articulated
    • acoustic streams that have more than one possible word boundary interpretation are sometimes called “oronyms”, e.g., “stuffy nose” vs. “stuff he knows”; “He saw the car go” vs. “He saw the cargo”; “I better do my laundry” vs. “I bet her five dollars.”
    • the linguistic context may be the only clue as the the correct perceptual interpretation
  • 2. Determination of the syntactical structure
    • we identify syllables and words, and recognize how these words combine to form linguistic clauses and these clauses combine to form a sentence; e.g., the listener must “parse” the input and determine the syntactic function of the incoming words
  • 3. Development of the propositional (conceptual) coherence
    • determination of the propositional content (“ideas”) represented and how these ideas are related
    • content from prior utterances and the new information are integrated
  • 4. Discourse comprehension
    • the full meaning is assembled not only from literal content but also on the basis of inference from information implied but not actually stated: “Asking the stranger for the time he ran quickly to make his appointment.”
      • stranger had a watch
      • stranger told the person the time
      • the questioner was late for the appointment
  • we look for and expect to find meaning in the utterances of others
• Top-Down Processing
  • Listeners develop useful expectations about probable word identity from real-world knowledge outside any information in a particular utterance
  • analysis proceeding both bottom-up and top-down simultaneously
  • spoken words in isolation are often totally unrecognizable (Pollack & Pickett, 1963)

He then considers the relationship between speech comprehension and working memory

• In contrast to reading, in speech perception there is no “looking back”
• This tends to results in shorter and syntactically simpler sentences that reflect the memory and processing constraints of the listener (There are other differences between spoken and written language)
• In the case of the perceptual clarification of ambiguous words, these must be held in some type of memory representation until the arrival of some “downstream” information makes it clear what that word must have been. Other types of memory constraints are reflected in so called “second-pass operations”, where we realize we have misunderstood a sentence as we originally heard it: “The old man the boats.” is an example of a “garden path” sentence:
  • “The old man the boats” [no, wait, that doesn’t make sense], “The old, man the boats” [yes, that makes sense]
• Wingfield suggests that substantial segments of speech input must be held in working memory when working with sentences that have embedded clauses, branching constructions, or need correct identification of referents: “The man who sold the car to the woman had red hair.” [Who had red hair? The man who sold the car to the woman]

Wingfield review sensory and cognitive changes in adult aging

• Age and auditory acuity
  • wide individual differences (a substantial proportion of the noninstitutionalized elderly have no speech problems with auditory acuity for speech), but: incidence of clinically significant hearing loss increases across the adult life-span
  • “presbycusis” literally means “old hearing” and has two important features:
    • aging differentially affects hearing at higher frequencies: affects the perception of high-frequency and low-energy sounds such as p, k, s, t sh, ch, and the voiceless th (as in “thin”)
    • “phonemic regression”: a lack of clarity for complex auditory signals such as speech; especially in presence of background noise
• Age and capacity for working memory
  • “working memory” refers to the cognitive system that includes both the ability to temporarily hold recently received information and a limited-capacity “computation space” that allows for the monitoring and manipulation of the information
  • working memory declines with age
• Age and speed of processing
  • slowing is the most consistent finding of the aging process
  • whether slowing causes inefficient processing or inefficient processing slows the system remains under investigation, but perceptual and response slowing are almost always shown in the aging literature (Salthouse, 2010)
  • older adults are especially sensitive to “time-compressed” speech
  • contextual effects can dramatically reduce age differences in recall of speech; older listeners make especially good use of the prosodic pattern (intonation pattern or pitch contour of speech, word stresses, variations in speech timing) of normal speech to help comprehension
  • slowing may help, but only if done in a considered way

Five principles to consider in regard to age and spoken language comprehension

1. Adult aging is often accompanied by declines in auditory acuity that may affect sensory processing of speech; however, linguistic knowledge remains preserved in older adulthood. This combination of loss and preservation leads to older adults making more use of “top-down” information drawn from linguistic context to supplement the impoverished bottom-up signal (this is also characteristic of the natural language processing of younger adults under difficulty listening conditions)
2. Slower processing rates and reductions in working memory can lead to overload (and error). Language structures that place a heavy burden on memory capacity may cause more difficulty (very long sentences, sentences that require referents that occurred far previously in the passage, sentence with high propositional density or complex syntax)
3. Normal prosody (intonation, timing, stress) can help older listeners; however, it is important to avoid exaggerated prosody to become patronizing “elderspeak”
4. It is important to avoid simple principles: “slower is better”–where one pauses is as important as how long one pauses. Pausing for a beat or two at the end of sentences, clauses, and main thought units will facilitate understanding and retention
5. Individual difference in regard to sensory and cognitive changes in adult aging can be very wide. We cannot design speech material for “the elderly”; we can design speech materials that respect processing limitations for adult listeners regardless of age: “These principles may take on more importance for some individuals than others, but they will be important to all.” (Wingfield, 1999, p. 224)