Audience Design and Social Relations in Aging

Abstract

This study asks two questions: (1) Do older adults modify their language based on age of the listener (audience design)? (2) Does social contact affect audience design in older adults? Older adults (n = 34; mean age = 82) engaged in an instructions task with two fictive listeners (a child and an adult) to test these questions. Results show that older adults used a greater total number of propositions and rapport-building devices and a lower type–token ratio when giving instructions to the child compared to the adult listener. Adults with more social interactions used more propositions when talking to a child. In addition, satisfaction with interactions was significantly positively related to task-tracking devices and negatively related to rapport-building devices by older adults. These results suggest that audience design and social relations are worth further study in language maintenance in older age.

Keywords

audience design social relations aging

Audience or recipient design is the ability of speakers to adjust their speech depending on the person with whom they interact. It has been observed both in spontaneous speech and in experimental contexts (Bell, 1984; Clark & Murphy, 1982; Hutchby, 1995). Audience design may be particularly sensitive to social contact because people who engage less with others (compared to those with more frequent interaction) may experience diminished capacity to tailor their messages to the needs of others.

Speakers typically vary their speech depending on real or perceived characteristics of different listeners, including age differences (e.g., young or old adults; Horton & Spieler, 2007) or different knowledge bases (e.g., nonnative speakers; Gass & Varonis, 1985). Investigations of audience design are typically examined in one of two ways: (1) changes in how members of a dyad modify their speech over the course of a conversation to reflect the development of common ground (Horton & Gerrig, 2002; Horton & Spieler, 2007; Hupet, Chantraine, & Nef, 1993) and (2) global assessments wherein a speaker makes an initial appraisal of another individual (e.g., as older) with particular cognitive and linguistic competencies and applies an internalized template of features, appropriate to talk with that category of individual, to their own speech (Adams, Smith, Pasupathi, & Vitolo, 2002; Giles & Ogay, 2009; Hummert, 1994).

However, findings are mixed regarding older adults’ ability to adjust their speech to different types of listeners. For instance, some older adults produce too much speech or speech that is off topic when talking to others (Arbuckle & Pushkar Gold, 1993; Pushkar Gold & Arbuckle, 1995). As such, these results suggest that older adults are insensitive to the needs of their interlocutors. Interestingly, Horton and Spieler (2007) found a different form of listener insensitivity in a referential communication task. Specifically, older and younger adults were asked to describe referents to familiar and unfamiliar conversational partners. When talking to familiar adults, older adults did not produce utterances displaying similarity with previous descriptions spoken to those familiars (which would have indicated sensitivity to audience design), whereas younger adults did make this adaptation. In contrast, Adams, Smith, Pasupathi, and Vitolo (2002) found that when older adults and college students were asked to retell a story to a child and to an adult experimenter, older adults differentiated the complexity of their speech to the child as compared to the adult more than did college students. The present study bridges the two types of research described previously by combining a task where failure to adapt one’s speech would lead to a lack of success, parallel to the referential communication study, but where listener assessments are based on an analysis of a listener as a member of a certain category (in this case a child or adult). Moreover, the present study employs an everyday practical task—a task that is familiar to older adults regardless of educational background—which should shed new light on older adults’ ability to achieve audience design.

Additionally, it is unknown whether older adults’ pattern of social relations might be related to their use of audience design. An individual who has a reduced frequency of interactions due to cognitive decline and/or loss of a partner might lose their sensitivity to the needs of different types of interlocutors. Their speech would then reflect less audience design, fewer adjustments, or different types of adjustments such as too little or too much information or fewer devices that help the listener track a set of instructions. Traditionally, social relations in aging are measured structurally, through network size and frequency of contact (Barnes, Mendes de Leon, Wilson, Bienias, & Evans, 2004; Holtzman et al., 2004; Seeman, Lusignolo, Albert, & Berkman, 2001; Zunzunegui, Alvarado, Del Ser, & Otero, 2003) or functionally (e.g., satisfaction with contact; Litwin, 1995; Wenger, 1997). The present study adopts and extends this framework for conceptualizing social relations by examining network size, frequency of contact, and satisfaction with contact in relation to audience design. Living arrangement was also included because those who live alone might be engaged in fewer interactions which would reduce their practice in designing their speech for others.

The central research questions were:

Research Question 1: Do older adults modify their speech when talking to speakers of different ages, in this case, a 10-year-old child and an adult?

Research Question 2: Are patterns of social relations (network size, frequency of interaction, or satisfaction with interactions) and living arrangement related to the way older adults talk to listeners of different ages?

Method

Participants

A community sample of 34 older adults was recruited via subject lists provided by a geriatrics center at a larger Midwestern R1 institution. Participants were sent recruitment letters with reply cards. If no response was received after approximately 10 days, one follow-up phone call was made to ascertain interest. Participants received US$20 for their effort and the study was approved by the Institutional Review Board of the R1 institution.

All participants self-identified as non-Hispanic White and the ages of participants ranged from 75 to 90 years (M_age = 82.30; SD = 4.10). Twenty-one participants identified as female and 13 identified as male. In terms of education level, 7 participants had less than or equal to a high school education, 27 had some college, an undergraduate degree, and some postgraduate education or a graduate/professional degree.

Materials and Stimuli

Demographic survey

Demographic data were collected as part of a larger study. For the purposes of this study, I included education which is well documented as associated with language behavior in aging (Connor, Spiro, Obler, & Albert, 2004; Hawkins & Bender, 2002; Neils et al., 1995). I also included two social variables, parent status and living arrangement. Because the adults in this study were asked to direct speech to a child, it was plausible that experience as a parent might lead parents to behave differently from nonparents (Campenni, 1999). Living arrangement was examined because evidence suggests that living alone is associated with an increased risk of dementia, which could impact language use (Bassuk, Glass, & Berkman, 1999; Fratiglioni, Hui-Xin, Ericsson, Maytan, & Winblad, 2000).

Social relations survey

To assess various aspects of social interaction, measures widely employed in the study of social relations were used (Fratiglioni, Paillard-Borg, & Winblad, 2004). Participants were asked to report the size of their network, how frequently they interacted with people (face-to-face, phone, letter, and email), and their satisfaction with the interactions.

Participants rated the frequency of their social interactions over the past 2 weeks on a 6-point Likert-type scale (1 = almost every day to 6 = never or hardly ever) and reverse coded for three different relationship categories: family, friends, and others. A social interaction score was created that included the summed frequency of interactions with each of the three groups of people, with higher scores reflecting more frequent social interaction.

Network size was assessed through a 4-point Likert-type scale (1 = none to 4 = more than 5). Participants were asked to report the total number of family, friends, and others with whom they interacted in the past 2 weeks. A network score was created that included the sum of the number of individuals in each of the three groups of people, with higher scores reflecting a larger network.

Satisfaction with social interaction was drawn from 4 items: “enjoyed talking,” “received new information,” “understood the way you feel,” and “got on my nerves” (reverse coded). Participants were asked to rate the extent to which they agreed with each item, on a 5-point Likert-type scale (1 = not at all to 5 = all the time). A satisfaction with social interaction score was created that included the summed satisfaction of the 4 items, with higher scores indicating greater satisfaction with social interactions.

Language tasks

I was interested in a task that would be familiar and suitable for both older men and women and which could appropriately be used with listeners of different ages (an adult and a child). For this purpose, I selected a sandwich-making task. Two types of sandwiches (grilled cheese and egg salad) were selected because they allowed for considerable variation in the range of detail provided to the listeners (e.g., contents of egg salad and how it is prepared, techniques for grilling cheese, and types of ingredients).

Age of the listener

The ages of the listeners were chosen with the following considerations in mind. First, I wanted to avoid child listeners who were too young because either they would not be able to perform the task or their age might elicit stereotyped “motherese” and hence show reduced variation in the variables of interest (Shatz & Gelman, 1973). An older child most likely would still require guidance by an adult to perform domestic tasks but could be expected to successfully complete making a sandwich. As for adult listeners, I selected a younger adult because this age should challenge the participants to think about what a 30-year-old might need to know, in contrast to an adult of the same age as the participant.

Two fictive listeners were created, a 10-year-old boy named Kenny and a 30-year-old named Harrison. Varying the age of the listener allowed me to tap into the differences in speech modifications based on with whom the participant was engaging in a task. The fictive listener technique was used successfully with older adults in previous research (Odato & Keller-Cohen, 2009). In that study, older and younger adults were shown pictures of fictive listeners differing by gender and age and were asked to make judgments about narratives attributed to the fictive listener. The participants were found to vary their responses depending on the age but not the gender of the fictive speaker. As such, the use of a photograph to represent the fictive listener was found to be useful when evaluating language behavior. Likewise, in the present study, a photograph of the fictive listener (either Kenny or Harrison) was placed on a stand and remained in prominent view throughout the testing and participants were told to direct their speech to that specific individual. Although speech production may differ with a non-fictive listener present (i.e., a confederate in the room), I expected that reliable differences based on the focal characteristic (age) of the listener would emerge.

Language Outcomes

In the present study, six linguistic features were examined in participants’ responses: (1) total number of propositions, (2) propositional density, (3) left-branching constructions as a proxy for syntactic complexity, (4) type–token ratio, (5) frequency of task-tracking devices, and (6) frequency of rapport-building devices. Number of propositions is a measure of the information content of the speech and propositional density roughly reflects how much information is packed into a sentence consistent with Kintsch (1974) and Turner & Greene (1977). Propositional density was calculated by dividing the number of propositions by the total number of words for a given participant. The Computerized Propositional Idea Density software (CPIDR; Brown, Snodgrass, Kemper, Herman, & Covington, 2008) was used for that purpose.

Left-branching constructions reflect syntactically more complex structures. Left-branching constructions, such as a very large rubber spatula pack in more information before the main noun (spatula) whereas man with a large umbrella presents the main noun first and then modifies it. Left-branching constructions were calculated by dividing the frequency of left-branching constructions by the total number of words. Psycholinguistic production and processing studies show left-branching constructions take longer to process and decrease in frequency in older age presumably to reduce demands on speech production. Some have argued that this decrease in use of left-branching construction is because this type of linguistic construction is cognitively taxing (Kemper, Thompson, & Marquis, 2001).

Type–token ratio (TTR), a measure of lexical variety in a speech sample, is the proportion of the number of different words out of the total number of words (Templin, 1957). Specifically, higher scores indicate greater variety of words. For instance, if each word used were different, the TTR would be 1; accordingly, if there were a great deal of repetition, the TTR would be low since there would be fewer different words.

In addition to these four language outcomes, two measures emerged from a review of the transcripts (Glaser & Strauss, 1967). The fifth measure I termed task-tracking devices, assesses instances in which the participants used phrases that enabled the listener to follow the instructions through the steps provided. Specifically, a composite score of task-tracking devices was created using the sum of the frequencies of the following pragmatic devices divided by the total number of words (for each listener and task):

Orientation: Orienting utterances appear in many communicative contexts (narratives, phone calls, and educational settings) and serve to engage the listener in the task at hand (Hopper, 1992; Labov & Waletsky, 1997). For instance, speech at the opening of the task that orients the child listener to the task at hand (e.g., Kenny, you tell me you’re hungry? Have you ever made your own lunch?).

Closings: Comments which indicate that the instructions are now complete and cue the end of an interaction (Sidnell, 2010), for example: And there you are, a perfect sandwich.

Step words: Words that enable the listener to keep track of the progression of actions and help provide a logical structure for instructions (e.g., first, next, and last).

Subroutines: Utterances that break down a particular step into substeps, enabling the listener to map out a plan for completion (Miller, Galanter, & Pribram, 1986). For instance, this might be an explanation of how to hard boil eggs: You put the pan in the sink and you run cold water for quite a while until the eggs chill.

I also coded transcripts for rapport-building devices (the sixth language measure) which is language used to establish a relationship with the listener through reference to joint action, concern, feedback, or probes to draw out the listener (Levinson, 1983). A composite score of rapport-building devices was created using the sum of the frequencies of the following pragmatic devices divided by the total number of words:

Collaboration: Words or phrases which indicate that the task is joint, that the speaker is “in there” with the listener: for example, We can do this; You can come over and we'll do it.

Feedback: Supportive speech aimed at helping the listener know how things are going, for example, You’re doing it just right; Good job.

Caution: Expressions of concern made to signal that the listener not injure himself or herself, for instance, Be careful because the stove is hot.

Questions: Inquiries into the listener’s understanding or experience, for example, Do you like white bread?

It could be argued that both task-tracking and rapport-building devices show a concern for the interlocutor, task tracking in helping the listener find their way through the task and rapport building, signaling connection with the listener. Use of these devices ought to be more common in talk to a child than an adult.

Hypotheses

I hypothesize that older adults will provide more information (propositions) to the child compared to the adult listener, given their appraisal of children as likely to lack the same world knowledge as adults. Propositional density will be lower and left-branching constructions will be less frequent in speech with a child listener compared to an adult due to their cognitive load. Type–token ratio should be lower when addressing a child, given the child’s more limited vocabulary as compared to an adult. Finally, older adults should produce more task-tracking devices when talking to a child as compared to an adult because of the child’s less well-developed cognitive skills; rapport-building devices should be more common with children as compared to adults because of the belief that children benefit from encouragement (Hattie & Timperley, 2007).

In terms of my hypotheses regarding social interaction, I predict that more socially engaged older adults should show more differentiation between adult and child listeners than those who are less engaged. They would produce more information, lower propositional density, less complex speech (i.e., fewer left-branching constructions), lower type–token ratio, and more task-tracking and rapport-building devices to children. However, when speaking with an adult, I predict that more socially engaged older adults should use a higher type–token ratio reflecting awareness of the listener’s greater vocabulary as compared to a child, and more complex speech including more left-branching structures and greater propositional density. Task-tracking and rapport-building devices may be more frequent in socially engaged adults whose social orientation might lead them to be concerned with listener success and support in task completion.

Procedure

Testing was administered on a large Midwest university campus in a quiet room (to avoid the effect of background noise) and the study took approximately 30–40 min to complete. Sessions were digitally audiorecorded to ensure the accuracy of the test administration and to create a digital record of the responses for the tasks. Tasks were administered by trained research assistants in the same order for each participant: language tasks, demographic survey, and social survey.

Participants were instructed:

I want you to explain to [name of listener] how to make a grilled cheese [egg salad] sandwich. He has never made a grilled cheese sandwich before so you will have to explain all the steps necessary for him to do this. After you’ve given [name of listener] the instructions, later today I’ll see how good a job he does making the sandwich so he’s depending on you to help him.

Each participant was first assigned to explain how to make an egg salad sandwich to either the child or adult (instructions for a grilled cheese sandwich were given to the other listener). There were no order effects regarding age of the listener across all of the dependent measures, ps > .09. In order to examine whether the type of sandwich (egg or cheese) participants were asked to explain affected the language outcomes, I conducted a series of t-tests, separated by listener (child or adult). Across all analyses for the child listener, there was no effect of type of sandwich on any of the language outcomes, t(32) = −1.75 to 1.19, ps > .09. Similarly, across the analyses related to the adult listener, there was no effect of type of sandwich, t(32) = −.90 to 1.72, ps > .10 (with the exception that those who explained how to make an egg sandwich used a significantly greater number of task-tracking devices than those who explained how to make a grilled cheese sandwich). This may be due to the number of steps associated with making an egg salad sandwich when an adult is addressed.

After the language task was finished, participants completed the demographic survey and social questionnaires. Before conducting analyses, all the transcripts were reviewed to assess whether any of the speech was off-topic because that would influence the interpretation of measures (e.g., number of words). No cases of off-topic speech were observed. The initial transcripts were prepared and then 15% re-transcribed with a reliability rate of 85%. Fifteen percent of all coding was recoded and achieved a reliability rate of 92%.

Results

Preliminary Analyses

A summary of the demographic, social, and linguistic measures appear in Table 1.

Table 1.

Descriptive Statistics: Demographic Social and Linguistic Variables.

	N	%	Range	Mean	SD
Age	34	—	75–89	82.26	4.10
Education			—	—	—
Some high school/high school graduate	7	21	—	—	—
Some college/college graduate and beyond	27	79	—	—	—
Living arrangement			—	—	—
Alone	13	39	—	—	—
With someone	20	61	—	—	—
Parent status			—	—	—
No children	6	18	—	—	—
Parent	28	82	—	—	—
Frequency of social interactions with network	34	—	16–36	27.62	4.60
Number in social network	34	—	13–24	20.09	2.90
Satisfaction with social interactions	34	—	29–60	48.56	6.42

First, I performed a series of t-tests to examine the relationship between the three demographic variables (level of education, living arrangement, and parent status) and the language measures of interest. Level of education (some high school/high school graduate vs. some college/college and beyond) was not significantly related to the number of propositions, propositional density, left-branching constructions, task-tracking devices, and rapport-building devices when speaking with the child, ps > .11. Older adults with lower levels of education (some high school/high school graduate) had higher type–token ratio scores when speaking with the child than those with higher levels of education, t(32) = 2.15, p = .04. In terms of the adult listener, level of education was not significantly related to proposition density, left-branching constructions, task-tracking devices, and rapport-building devices, ps > .13. However, when speaking to the adult, older adults with lower levels of education used fewer propositions and had higher type–token ratio scores than older adults with more education, t(32) = −2.60, p = .01 and t(32) = 2.16, p = .04, respectively. For both listener conditions, living arrangement (alone vs. with someone) and parent status (not a parent vs. parent) were not significantly related to any of the six language outcomes, ps > .17.

In sum, these three demographic variables were largely not related to the language outcomes of interest. The pattern of results for level of education and type–token ratio for both listeners departed from this. It may be the case that those with lower education are less sensitive to lexical variation as a social tool. This may explain why their type–token ratio was higher than that of more well-educated adults for both the child and adult listeners. Moreover, living arrangement and parent status were not related to any of the language outcomes. Given the null pattern of results, these three demographic variables were not included as covariates in subsequent analyses.

Do Older Adults Modify Their Speech When Speaking to a Child Versus an Adult?

To test the present study’s first hypothesis, I performed six paired t-tests to compare the age of the listener for the sandwich-making task on the use of total number of propositions, propositional density, type–token ratio, left-branching constructions, and rapport-building devices. As expected, older adults modified their speech in certain ways when speaking to a child. Specifically, older adults used a greater total number of propositions, and rapport-building devices when giving constructions to the child compared to the adult listener (see Table 2). Moreover, older adults had higher type–token scores when speaking with the adult compared to the child, indicating they were sensitive to the diversity of vocabulary their listener was likely to possess. Inconsistent with my hypotheses, older adults did not differ in their use of propositional density, left-branching constructions, and task-tracking devices when speaking to a child compared to an adult.

Table 2.

Paired t-Test Analyses: Language Outcomes for Child and Adult Listeners.

	Listener
Scales	Range	Child	Range	Adult	t
Scales	Range	M (SD)	Range	M (SD)	t
Total number of propositions	46–601	162.74 (118.61)	40–256	110.18 (57.05)	3.82***
Propositional density	0.43–0.59	0.53 (0.04)	0.46–0.63	0.53 (0.04)	0.29
Left branching	0.00–0.08	0.02 (0.01)	0.00–0.03	0.02 (0.01)	1.32
Type–token ratio	0.21–0.70	0.42 (0.10)	0.34–0.64	0.48 (0.08)	−4.90***
Task-tracking devices	0.02–0.10	0.05 (0.02)	0.01–0.12	0.05 (0.02)	0.50
Rapport-building devices	0.00–0.08	0.02 (0.02)	0.00–0.04	0.01 (0.01)	2.07*

Note. *p < .05. **p < .01. ***p < .001.

Are Social Relations Related to Patterns of Speech Among Older Adults?

Given the exploratory nature of the present study, I wanted to examine the relationships between three different measurements of social relations and the six language outcomes. Thus, to test the present study’s second hypothesis, I conducted bivariate correlations between the measures of social relations and language outcomes (see Table 3). Given that the demographic variables of interest were not significantly related to the language outcomes, bivariate correlations were performed, as appropriate for the sample size. See Table 3 for interitem correlations between the language variables associated with the child and adult listeners.

Table 3.

Bivariate Correlations Among Social Interaction Variables and Language Outcomes by Listener.

	1	2	3	4	5	6	7	8	9
Frequency of social interactions	—	—	—	.32^†	.26	.06	−.21	−.18	.06
Network size	.58***	—	—	.11	.16	.04	−.13	.21	.11
Satisfaction with interactions	.24	.48**	—	−.07	−.02	.06	−.01	.01	.03
Total number of propositions	.36*	.13	−.15	—	.29^†	.25	−.84***	−.27	−.20
Propositional density	.06	−.07	−.03	.17	—	−.10	−.04	.01	.05
Left branching	−.21	.03	.30^†	−.30^†	−.11	—	−.35*	.30^†	.14
Type–token ratio	−.28	−.05	.10	−.84***	−.07	.19	—	.14	.25
Task-tracking devices	−.01	.17	.37*	−.25	−.18	.52**	.45**	—	.36*
Rapport-building devices	.01	.03	−.38*	.26	−.09	−.22	−.19	−.10	—

Note. Values above the diagonal report correlations for the adult listener (shaded) and below report correlations for the child listener.

^† p > .05 to < .10. *p < .05. **p < .01. ***p < .001.

For the instructions task with the child audience, only one of the correlations between frequency of social interaction and the six language outcomes was significant. Specifically, greater frequency of social interaction was significantly related to the use of more propositions when speaking to the child (ps < .05). Inconsistent with the present study’s hypotheses, frequency of social interaction was not significantly related to propositional density, type–token ratio, left-branching constructions, task-tracking devices, and rapport-building devices. However, consistent with the study’s hypotheses, satisfaction with social interactions was significantly positively related to task-tracking devices and negatively related to rapport-building devices. Greater satisfaction with social interactions was related to greater use of left-branching constructions at the level of a trend. Network size was not significantly related to any of the six language outcomes associated with the child listener, indicating that frequency of social interaction and satisfaction with social interactions may be more nuanced measurements of social interaction.

In terms of the instructions task with an adult listener, frequency of social interaction was not significantly related to the six language outcomes (see Table 3). Although the correlations were not statistically significant, one correlation trended toward significance. Higher frequency of social interaction was marginally related to greater use of propositions (p = .06). However, network size and satisfaction with social interactions were not significantly related to any of the six language outcomes associated with the adult listener.

Taken together, frequency of social interaction was linked with two of the language outcomes associated with the child audience; specifically, total number of propositions and task-tracking devices. Social interaction was not related to speech patterns among older adults giving instructions to an adult listener. Moreover, network size and satisfaction with social interactions were not significantly related to any of the language outcomes.

Discussion

The present study explored whether older adults modify their speech when speaking to a child and adult listener and whether the nature of older adults’ social contact was related to the way they spoke to listeners of different ages.

Audience Design

Previous research has offered conflicting views of the ability of older adults to modify their speech to listeners of different ages (Adams et al., 2002; Horton & Spieler, 2007). The results of the present study suggest that when older adults are asked to perform an everyday task, they can display sensitivity to listeners of different ages. Consistent with the study’s hypotheses, older adults were found to use a greater total number of propositions and rapport-building devices and a lower type–token ratio when giving instructions to the child compared to an adult. These results indicate that older adults vary their speech to different aged listeners by providing more information as well as expressing rapport more often for child listeners than adult listeners and using greater lexical diversity with adults. Thus, there was differentiation between the two types of listeners. If we think of speakers as having particular repertoires used with different types of listeners, then the choice of not to use a particular device or construction differently with various types of listeners is not a case of no difference but rather a choice to use that device similarly. That is, the repertoires may overlap in some patterns of use and differ in others, related to Bell’s (2007) concept of style. In future research, it would be very useful to look at modifications to three types of listeners, for example, older adult peers as compared to two younger groups, in order to develop a more nuanced picture of speech adjustments.

The absence of difference between an adult and a child listener in propositional density and the use of left-branching constructions used may be due to the instructions genre. Although modern recipes are typically written, their counterpart, oral recipes are a procedural genre characterized by a high degree of formulaic language. Thus, it may be that the everyday instructions genre has some optimal level of information density and sentence complexity toward which all speakers aim. This may likewise be true of task-tracking devices. The present study focused on a task genre with formulaic features; therefore, it would be of importance for future research to systematically compare the language used in different genres when interacting with speakers of different ages to evaluate whether listener sensitivity is affected by the type of talk.

Social Relations and Language

Previous work has documented the relationship between social relations and word retrieval in aging (Keller-Cohen, Fiori, Toler, & Bybee, 2006); however, the present study is the first to examine how older adults’ engagement in social relations is related to their everyday use of language. Building on previous work, we expected that more socially engaged participants (more frequent interactions, a larger network, and greater satisfaction with interactions) would demonstrate greater differentiation in their speech when talking to listeners of two different ages. Although network size was not found to be related to the language measures of interest, individuals who reported more frequent social interactions provided a fictive child listener with more information (more propositions). Those with higher satisfaction with interactions used more task-tracking devices and fewer rapport-building devices when talking to a child; satisfaction was marginally related to left-branching constructions. In instructions to an adult, the relationship between number of propositions and frequency of social interactions was marginally significant, a parallel but weaker relationship than speech to the child. The greater information in response to the child could be an assessment of the listener’s knowledge state, a sensitivity that could grow out of regular contact with others. The parallel although weaker finding in talk to the adult listener might also indicate that providing more information is also an indication of sociability. Future work would do well to probe sociability and its relation to patterns of talk in audience design.

Understanding the relationship between satisfaction with interactions and speech to a listener will require more attention in the future. Satisfaction with interactions is one way of talking about the quality of verbal exchanges with others. Satisfaction might lead an individual to feel they don’t need to make special effort in interactions because things will work out well or it might stimulate certain types of communicative behavior because those patterns have been rewarding. Perhaps those who feel their interactions are largely satisfying might not experience a need to build rapport with unfamiliar others and could account for the pattern observed here. The explanation for higher rates of satisfaction associated with higher rate of task-tracking devices to the child listener in contrast could grow out of the satisfaction derived from helping another verbally. In either case, the findings make satisfaction worth further pursuit in studies of communication in older age.

Given the link between social interaction and language production, I next consider what types of social interactions might be of greatest benefit to an older adult for maintaining language skills. One possibility is that frequent interactions with a smaller range of particular individuals, such as members of one’s social network, may be of greatest benefit to older adults. Specifically, frequent contact with close associates may elicit a great amount of speech production because of the older adult’s comfort with the interlocutors. Another possibility is that frequent interactions with a wider range of people representing a more diverse set of membership categories may be beneficial for maintaining language production among older adults. For instance, it could be argued that a more diverse range of interlocutors would be more beneficial for continued activation of language skills in aging than equally frequent interactions with a more limited range. The linguistic assessments a speaker makes in order to interact with different types of individuals ought to tap into a broader set from one’s linguistic repertoire, keeping active a wider range of one’s linguistic resources. In fact, Keller-Cohen, Fiori, Toler, and Bybee (2006) found that older adults who interacted with a higher proportion of friends and a more diverse range of individuals performed better on the Boston Naming Test (a measure of word retrieval) than those whose interactions were weighted more toward family members.

How might the frequency of social interactions be related to providing instructions in particular and the maintenance of language skills more generally? First, it should be noted that the current data illuminate a relationship between some aspects of language use and the frequency of social interactions at a point in time. Given the limitations of the data set, it is unknown whether the frequency of interactions over the course of time is related to listener-tailored messages or some other factor, such as the preservation of cognitive skills, enables individuals to engage in more social interactions that are important in listener-sensitive messages. Potentially, social relations may moderate the relationship between cognition and language. However, it might also be the case that some of these relationships are reciprocal. Small, Dixon, McArdle, and Grimm (2012) found that social engagement may affect and be affected by cognitive activities, depending upon the cognitive domain (e.g., verbal speed, episodic memory, and semantic memory). Thus, future research should consider whether social interaction moderates the relationship between cognition and language production.

Finally, the findings here are consistent with research that shows that audience design is a characteristic of talk that occurs not only in older age, as this and other studies have shown but also appears across the life span (Bell, 1984; Clark & Murphy, 1982; Shatz & Gelman, 1973). It seems unlikely that one’s social world would play a role in audience design only in older age although this study cannot speak directly to that. It might be that some aspects of social relations need to decrease to a tipping point where they negatively impact language and other cognitive behavior. This is certainly an avenue worth exploring in future work.

Limitations and Strengths

A strength of the current research is the fictive listener, a technique that avoids the effect of live listeners who vary considerably in interactional style and level of interest in the task. The success of this method is consistent with previous research (Odato & Keller-Cohen, 2009). The older adult participants in the present study engaged fully in interaction with the fictive child and adult, accepting their fictive identity saying such things as Why Harrison, you’re 30 years old and you don’t know how to make a grilled cheese sandwich? or Kenny, I know you’re going to enjoy this. Although the speech produced in this case may differ in some ways from speech one would observe in a live face-to-face interaction, the speakers did produce different types of language in response to the two fictive listeners. This suggests that older adults were able to imagine each listener and craft tailored responses based on the age of the listener. This approach is certainly worthy of further consideration in tasks involving audience design with co-participants of different ages. At the same time, it would be interesting to compare language produced in response to a fictive and live listener to better understand the possibilities and limitations of the fictive listener technique.

An important limitation of the study is a sample size sufficient to simultaneously examine age of the listener, demographics, and social interaction variables. A larger sample would be needed to evaluate the role each of these variables plays in how older adults modify their speech to different types of listeners as well as how different profiles of social contact are related to these modifications. In addition, a range of language tasks would be important to employ in future research, in order to examine how the function of different types of talk affects audience design. It would also be useful to explore a range of listeners who differ along other dimensions such as gender and ethnicity to assess the range of competencies older adults carry with them as they age. Finally, the social measures employed in the present study were exploratory; thus, future research is needed to refine and expand these measures. Taken together, consideration of these limitations will provide future researchers with a more nuanced view of the relationship between language and social behavior in aging.

Conclusion

The ability to adjust one’s speech to a particular listener is significant in many domains of an older adult’s life. Sensitivity to what an interlocutor knows (or is likely to know) and what they are likely to find relevant, interesting, or valuable, positions a speaker as an attentive and interested conversational partner. A speaker who provides inadequate information, off-topic details, or excessive information is difficult for others to understand and less likely for others to show interest in engaging (Hamilton, 1994; Rusher & Hurley, 2000). For instance, an appointment with a physician is an important context to examine audience design. Patients who cannot provide an intelligible narrative for their health care provider may not elicit the most interested behavior by that provider; thus, they may not receive the level of care needed. Similarly, consider older adults interactions around written materials. Older adults are confronted with a range of written materials about their health care that often warrants conversations with others about that information. Anticipating what another is likely to know and characterizing one’s own insurance situation are fundamental elements to obtaining the information the older adult needs. Framing one’s needs in the consumer arena is also impacted by anticipating what the potential audience knows and can provide. Regular practice at anticipating the interactional needs of others may provide older adults with opportunities to assess what others know and to hone one’s message formation skills. The results of this study suggest that older adults are able to modify their speech in different ways to a child versus an adult which is a first step in exploring what speech modifications older adults are able to make to different types of listeners. Moreover, sensitivity to a listener’s needs is an important survival skill for navigating the different interlocutors older adults encounter. Finally, the results of the present study provide some evidence that the role of social contact in maintaining one’s linguistic repertoire in later adulthood is a fruitful avenue for future exploration.

Footnotes

Acknowledgments

I want to thank the Claude D. Pepper Center, University of Michigan, funded through grant AG024824 from the National Institute of Health for its role in providing research participant candidates. I also wish to acknowledge the help of Amy Moors, Keith Rainwater, Laura Miller, Erica Beck, Lauren Squires, David Medeiros, Terry Szymanski, Heather Taylor, Julia Seng, Jessica McKeever, Lucy Yang, Lesley Materne, and Kathy Welch for their contributions to data collection and analysis and the lab of Holly Craig for transcription. I also appreciate the valuable advice provided by three anonymous reviewers of this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Adams

Smith

Pasupathi

Vitolo

(2002). Social context effects on story recall in older and younger women: Does the listener make a difference? Journals of Gerontology: Psychological Sciences and Social Sciences, 57, 28–40.

Arbuckle

T. Y.

Pushkar Gold

(1993). Aging, inhibition, and verbosity. Journals of Gerontology, 48, P225–P232.

Barnes

Mendes de Leon

Wilson

Bienias

Evans

(2004). Social resources and cognitive decline in a population of older African Americans and whites. Neurology, 63, 2322–2326.

Bassuk

Glass

Berkman

(1999). Social disengagement and incident cognitive decline in community-dwelling elderly persons. Annals of Internal Medicine, 131, 165–173.

Bell

(1984). Language style as audience design. Language in Society, 13, 145–204.

Bell

(2007). Style and the linguistic repertoire. In Llamas

Mullany

Stockwell

(Eds.), The Routledge companion to sociolinguistics (pp. 95–100). London, England: Routledge.

Brown

Snodgrass

Kemper

S. J.

Herman

Covington

(2008). Automatic measurement of propositional idea density from part-of-speech tagging. Behavior Research Methods, 40, 540–545.

Campenni

(1999). Gender stereotyping of children’s toys: A comparison of parents and nonparents. Sex Roles, 40, 121–138.

Clark

Murphy

(1982). Audience design in meaning and reference. In Ny

J.F. Le

Kinstsch

(Ed.), Language and comprehension (pp. 287–299). New York, NY: North-Holland.

10.

Connor

Spiro

Obler

Albert

(2004). Change in object naming ability during adulthood. Journal of Gerontology: Psychological Sciences, 59, 203–209.

11.

Fratiglioni

Hui-Xin

Ericsson

Maytan

Winblad

(2000). Influence of social network on occurrence of dementia: a community-based longitudinal study. The Lancet, 355, 1315–1319.

12.

Fratiglioni

Paillard-Borg

Winblad

(2004). An active and socially integrated lifestyle in late life might protect against dementia. The Lancet, 3, 343–353.

13.

Gass

S. M.

Varonis

(1985). Variation in native speaker speech modification to non-native speakers. Studies in Second Language Acquisition, 7, 37–58.

14.

Giles

Ogay

(2009). Communication accommodation theory. In Whaley

B. B.

Samter

(Eds.), Explaining communication: Contemporary theories and exemplars (pp. 325–344). Mahwah, NJ: Lawrence Erlbaum.

15.

Glaser

Strauss

(1967). The discover of grounded theory: Strategies for qualitative research. New York, NY: Aldine.

16.

Hamilton

(1994). Conversations with an Alzheimer's patient. Cambridge, MA: Cambridge.

17.

Hattie

Timperley

(2007). The power of feedback. Review of Educational Research, 77(1), 81–112.

18.

Hawkins

Bender

(2002). Norms and the relationship of Boston Naming Test performance to vocabulary and education: A review. Aphasiology, 16, 1143–1153.

19.

Holtzman

Rebok

Saczynski

Kouzis

Doyle

Eaton

(2004). Social network characteristics and cognition in middle-aged and older adults. The Journals of Gerontology Series B: Psychological and Social Sciences, 59(6), P278–P284.

20.

Hopper

(1992). Telephone conversation. Bloomington, IN: Indiana University Press.

21.

Horton

W. S.

Gerrig

R. J.

(2002). Speakers' experiences and audience design: knowing when and knowing how to adjust utterances to addressees. Journal of Memory and Language, 47, 589–606.

22.

Horton

W. S.

Spieler

(2007). Age-related differences in communication and audience design. Psychology and Aging, 22, 281–290.

23.

Hummert

M. L.

(1994). Stereotypes of the elderly and patronizing speech. In Hummert

M. L.

Wiemann

J. M.

Nussbaum

J. F.

(Eds.), Interpersonal communication in older adulthood: Interdisciplinary theory and research (pp. 163–184). Thousand Oaks, CA: Sage.

24.

Hupet

Chantraine

Nef

(1993). References in conversation between young and old normal adults. Psychology and Aging, 8, 339–346.

25.

Hutchby

(1995). Aspects of recipient design in expert advice-giving on call-in radio. Discourse Processes, 19, 219–238.

26.

Keller-Cohen

Fiori

Toler

Bybee

(2006). Social relations, language and cognition in the ‘oldest old'. Ageing and Society, 26, 585–605.

27.

Kemper

Thompson

Marquis

(2001). Longitudinal change in language production: Effects of aging and dementia on grammatical complexity and propositional content. Psychology and Aging, 16, 600–614.

28.

Kintsch

(1974). The representation of meaning in memory. Hillsdale, NJ: Erlbaum.

29.

Labov

Waletsky

(1997). Narrative analysis: Oral versions of personal experience. Journal of Narrative and Life History, 7, 3–38.

30.

Levinson

(1983). Pragmatics. Cambridge, MA: Cambridge University Press.

31.

Litwin

(1995). The social networks of elderly immigrants: An analytic typology. Journal of Aging Studies, 9, 155–174.

32.

Miller

Galanter

Pribram

(1986). Plans and the structure of behavior. New York, NY: Adams Bannister Cox.

33.

Neils

Baris

J. M.

Carter

Dell'aira

A. L.

Nordloh

Weiler

Weisiger

(1995). Effects of age, education, and living environment on Boston Naming Test performance. Journal of Speech, Language, and Hearing Research, 38, 1143–1149.

34.

Odato

Keller-Cohen

(2009). Evaluating the speech of younger and older adults: Age, gender and speech situation. Journal of Language and Social Psychology, 28, 457–475.

35.

Pushkar Gold

Arbuckle

(1995). A longitudinal study of off-target verbosity. Journals of Gerontology: Psychological Sciences and Social Sciences, 50, P307–P315.

36.

Rusher

J. B.

Hurley

M. M.

(2000). Off-target verbosity evokes negative stereotypes of older adults. Journal of Language and Social Psychology, 19, 141–149.

37.

Seeman

Lusignolo

Albert

Berkman

(2001). Social relationships, social support and patterns of cognitive aging in healthy, high-functioning older adults: MacArthur studies of successful aging. Health Psychology, 20(4), 243–255.

38.

Shatz

Gelman

(1973). The development of communication skills: Modifications in the speech of young children as a function of listener. Monographs of the Society for Research in Child Development, 38, 1–38.

39.

Sidnell

(2010). Conversation analysis: An introduction. Malden, MA: Wiley-Blackwell.

40.

Small

Dixon

McArdle

Grimm

(2012). Do changes in lifestyle engagement moderate cognitive decline in normal aging? Evidence from the Victoria Longitudinal Study. Neuropsychology, 26, 144–155.

41.

Templin

M. C.

(1957). Certain language skills in children: Their development and interrelationships. Minneapolis: University of Minnesota Press.

42.

Turner

Greene

(1977). The construction and use of a propositional text base. Boulder: Institute for the Study of Intellectual Behavior, University of Colorado.

43.

Wenger

G. C.

(1997). Social networks and the prediction of elderly at risk. Ageing and Mental Health, 1, 311–320.

44.

Zunzunegui

M. C.

Alvarado

Del Ser

Otero

(2003). Social networks, social integration and social engagement determine cognitive decline in community-dwelling Spanish older adults. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 58B(2), S93–S100.