Active Is as Active Does: Deliberative and Non-deliberative Political Communication in Context

The development of theories of citizens’ political communication has been hindered by a debate over whether such communication is best characterized as deliberative or non-deliberative. This article aims to overcome that impasse with a new account of citizens’ political communication informed by theories of message production and sense-making: the goals–sense-making–justification (GSJ) model. This model holds that citizens’ political-communicative behavior is influenced by multiple goals and cognitive plans, which generally vary in different contexts. This variation helps to explain why citizens’ informal political discussions during non-election periods rarely feature reason-giving—and so can be understood as non-deliberative—whereas such discussions during major-election campaigns often feature the reason-giving that is characteristic of deliberation. Empirical results—from an original study and previous research— are consistent with several features of the model, but findings concerning persuasive and advice-giving goals and the use of information-seeking behaviors are mixed, and more evidence is needed of the role of cognitive plans in citizens’ political communication.

Political-communication theorists highlight citizens' agency (Cho et al. 2009), yet disagree about that agency's nature. Some argue citizens' everyday political discussion is non-deliberative (e.g., Eveland et al. 2011), whereas others cast deliberation as central to such discussion (e.g., Neblo 2015). We reconcile this conflict with a new account of citizens' political communication, the goals-sensemaking-justification (GSJ) model. Under this model, different contexts cause different goals to shape citizens' political discussion, goals which activate in citizens' minds different communicative plans, which encourage different communicative behaviors. These relationships among contexts, goals, plans, and behaviors help explain the generally non-deliberative character of citizens' informal political communication outside of major political events and the deliberative character of much of that communication during such events. Moreover, the model aims to accord with empirical evidence about most citizens' political talk and to explain the capacity of citizens, without argumentation training, to engage competently in reason-giving during formal deliberations (Collingwood & Reedy 2012). While delineating the model, we propose hypotheses and then present preliminary evidence to seed a further research agenda.

Citizens' Political Discussion
Political-communication 1 scholars disagree about whether democratic-deliberative theories account for citizens' 2 actual political-communicative practices. We define democratic deliberation as interpersonal communication, characterized by reason-giving, about public policy 3 or politics 4 (Gastil 2008;Habermas 1996). To be sure, theorists highlight multiple deliberative criteria, including equal and respectful treatment, and achieving mutual understanding, the absence of polarization, and procedural legitimacy (Carcasson & Sprain 2015;Fishkin 2009;Gastil 2008). Nonetheless, the GSJ model defines deliberation in terms of reason-giving on three grounds.
Among accounts characterizing citizens' political talk as deliberative are theories explaining deliberative practices' communication. This model explains the deliberative or non-deliberative nature of such communication as arising from associations among contexts, goals, and plans.
Before presenting the GSJ model, we note that the model's theoretical sources were designed to account for face-to-face, online, and hybrid in-person and online settings. 6 Consequently, the following account of the GSJ model does not differentiate between face-to-face and online interactions. 7

Goals-Sense-making-Justification Model
In presenting the GSJ model, we first explain the sensemaking (Dervin 1989) 8 and goals-plans-actions (GPA) frameworks (Dillard 2015). Under sense-making theory, individuals generally employ information to pursue goals (Dervin 1989). Faced with a political decision, most democratic citizens, possessing little political knowledge (e.g., Delli Carpini & Keeter 1996), experience information needs, generating uncomfortable feelings of uncertainty, disorientation, and anxiety (Kuhlthau 1991). To alleviate that discomfort, citizens seek new information to satisfy those needs (Kuhlthau 1991). By acquiring new information required to make the decision, citizens achieve greater orientation, understanding, and confidence (Dervin & Frenette 2001). Moreover, personality traits affect information-seeking. Heinström (2006) found neuroticism was positively associated with quick searching of few sources, extraversion and openness with broad searching of many sources, and openness and conscientiousness with deep searching.
In addition to sense-making theory, the GSJ model draws from the GPA model, holding that message production begins with the context, 9 a pattern of social events in which individuals perceive ' obstacles and opportunities' (Dillard & Solomon 2000: 170). Each goal resides in memory, with ' cognitive rules' specifying particular opportunities and obstacles of contexts, which trigger goal activation (Wilson 1990: 81-82). Citizens' perception of a context as presenting particular opportunities and obstacles leads cognitive rules to activate particular goals 10 of two kinds. Primary goals launch and define communicative interactions, while secondary goals limit communicative conduct (Dillard et al. 1989). Thus citizens pursue multiple goals during political communication: making a political decision; maintaining relationships, reputation, and selfimage; ' obtain [ing] information;' 'provid [ing] guidance;' ' chang[ing others'] opinion' (Kellermann 2004: 407); 'selfdevelopment' (Conover et al. 2002: 52); 'pass[ing] the time; ' and 'generat[ing] an interesting discussion (Eveland et al. 2011(Eveland et al. : 1091. 11 Those goals activate plans: sets of cognitive representations of communicative behaviors intended to advance the goals (Berger 1997). 12 Plans are used because they help predict and overcome obstacles, increasing the likelihood of goal attainment, and facilitate 'interpersonal coordination' when disclosed (Berger 1997: 24). Once goals are stimulated, the mind, to conserve limited cognitive resources, searches memory for already-formulated plans depicting behaviors likely to achieve those goals and retrieves an appropriate plan if available. If no such plan is found, a suitable plan is constructed ' on the fly' (Berger 1997: 26-28). Moreover, traits of self-monitoring and cognitive complexity are positively associated with plan complexity (Berger 1997: 42-46).
Once activated, a plan spurs performance of behaviors depicted in it (Dillard 2015). 13 Plans include instructions to the sensorimotor system to perform depicted behaviors (Greene 1984). Stored in memory, plans are repeatedly retrieved and revised for re-use in particular contexts to conserve limited mental resources (Berger 1997). 14 Thus different goals generally define (i.e., function as primary goals in) communicative interactions in different contexts, different plans are generally activated in different contexts, and communicative conduct usually varies between contexts. Yet in distinct but similar contexts, individuals perceive similar opportunities and obstacles, which cause similar goals to define interactions, which activate similar plans, which foster similar behaviors.
The GSJ model posits four contexts of citizens' political communication: informal discussion in one's close social network of family and friends (Minozzi et al. 2020) outside of a major political event (Informal Ordinary Context) and during such an event (e.g., Cho et al. 2009) (Informal Major-Political-Event Context) and formal discussion with strangers (e.g., in a jury) (Sprain & Gastil 2013) outside of such an event (Formal Ordinary Context) and during such an event (Formal Major-Political-Event Context).

Informal Ordinary Context
In the Informal Ordinary Context, no major political event, such as a major election, 15 exerts pressure on most citizens to make a political decision. Citizens generally discuss politics in casual conversations with family (Warner & Colaner 2016) and friends (Minozzi et al. 2020). These discussions are generally infrequent, interwoven with small talk, and of uncertain regularity (Eveland et al. 2011;Goldsmith & Baxter 1996). In this context, citizens generally perceive familial and friendship relationships as most salient and as opportunities for pleasant, relaxed interaction (Dillard & Solomon 2000: 170). Further, because in this context most citizens feel no social pressure to make a political decision, citizens experience no need for decision-related information. Perceiving these features of the context likely leads cognitive rules (Wilson 1990) to activate social and personal goals, which should 16 generally define these interactions (Conover et al. 2002;Eveland et al. 2011) (Table 1): H1(a). Goals of maintaining relationships, reputation, and self-image; passing the time; self-development; and stimulating interesting discussion will define citizens' interactions to a greater extent than goals of making a political decision, obtaining information, providing guidance/advice, or changing others' opinions.
Those primary goals should activate plans, because-in this and other contexts-plan use facilitates goal attainment and interpersonal coordination 17 (Berger 1997: 24). Given limited cognitive resources, among individuals experienced in small talk, already-formed conversational plans should be retrieved from memory; whereas, among those lacking such experience, the mind is likely to construct a plan then and there (Berger 1997: 26-28). The plan should contain representations of behaviors involving exchanging political messages, without reasons, like tokens during small talk. Accordingly, in political talk with family and friends in the Informal Ordinary Context H1(b). Citizens will plan to use (i) information-seeking and (ii) reason-giving behaviors infrequently.
Because citizens are unlikely to devote cognitive effort to casual conversations, plans for such conversations will probably lack complexity, and so in that context, personality traits are unlikely to influence plan complexity (cf. Berger 1997: 42-46).
The activated plan should prompt performance of behaviors depicted in it-in this and other contextsthrough instructions, encoded in the plan, to the sensorimotor system 18 (Greene 1984). Thus, in political discussions with family and friends in the Informal Ordinary Context H1(c). Citizens will infrequently use (i) information-seeking and (ii) reason-giving behaviors. 19

Informal Major-Political-Event Context
Informal political talk among family and friends 20 also occurs during major political events: salient, politically relevant phenomena that capture public attention and lead most citizens to experience social pressure to make political decisions, such as whether to protest (Curato 2019) or how to vote (Coleman 2012). Major electionsmajor political events involving electoral decisions 21 -are emphasized herein, because they are the best-documented of those events.
In the Informal Major-Political-Event Context, the GSJ model predicts two phases of goal activation. The first concerns information-seeking ( Table 1). As the political decision looms, citizens experience pressure to decide from media and citizens' close-social-network members (Coleman 2012). 22 That pressure should increase the decision's salience, presenting opportunities to exercise freedom of choice and to placate those exerting pressure (Dillard & Solomon 2000: 170). Perceiving those opportunities should cause cognitive rules to retrieve, in the citizen's mind, the decision-making goal as a primary goal (Wilson 1990: 81-82).
Activating that goal should raise citizens' awareness that they lack information. First, many citizens lack sufficient information to make the political choice, including information about political parties' basic positions (Delli Carpini & Keeter 1996). In a major election, without seeking information, such citizens cannot make use of voting cues, such as party labels, that foster political polarization (Sides et al. 2018). 23 Many citizens also need information concerning justification. Because the citizen lives in a close social network whose members monitor his/her public political Note. The goal called Obtain information includes the goal of gathering knowledge. In the Informal Major-Political-Event Context, the personality trait of neuroticism is expected to be negatively associated, and the personality traits of extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity are expected to be positively associated, with the complexity of information-seeking plans and with the performance of information-seeking behaviors; information-seeking-plan complexity is expected to mediate partially associations between traits and performance. In the Informal Major-Political-Event Context, the personality traits of self-monitoring and cognitive complexity are expected to be positively associated with the complexity of justification plans and with the performance of reason-giving behaviors; justification-plan complexity is expected to mediate partially associations between traits and performance. See text for details.
decisions (Coleman 2012), the citizen, in informal political talk with those members, should seek to fulfill relationaland reputational-maintenance goals as secondary goals (Dillard 2015). Consequently, the citizen will probably plan for future discussions with those members (Dillard 2004), in which the citizen will need to justify his/her decision or seek to persuade members to change their decisions, with reasons acceptable to those members. Fashioning those justifications requires information about those members' political views and values. Yet in the Informal Ordinary Context, political talk is generally thin on reasons (Eveland et al. 2011) that might disclose those values and views. Consequently, in the Informal Major-Political-Event context, the citizen should realize they lack sufficient information about those members' political views and values to formulate justifications acceptable to those members. Those information needs should generate emotions of uncertainty, disorientation, and anxiety (Kuhlthau 1991); emotions are therefore central to informal deliberation (Neblo 2020). To resolve those feelings, the citizen should seek new information to satisfy those needs (Kuhlthau 1991) 24 and should perceive those needs as obstacles to decision-making (Dillard & Solomon 2000: 170). That perception should cause cognitive rules to activate the information-seeking goal (Wilson 1990: 81-82), which should join the decision-making goal in defining citizens' political-communicative interactions.
Activating the information-seeking goal should stimulate planning (Berger 1997: 24). Possessing limited mental resources, citizens with experience gathering information about decisions or about family and friends' views should retrieve from memory an already-existing plan (Berger 1997: 26), whereas in the minds of citizens lacking that experience a plan is likely to be developed (Berger 1997: 26-28). That plan is expected to contain representations of behaviors for acquiring information about political choices and about close-social-network members' political views and values (Kuhlthau 1991).
Personality traits will probably influence citizens' information-seeking plans. Neuroticism, linked to quick searching of few resources (Heinström 2006), is likely to be negatively associated with the complexity of those plans. Conversely, because extraversion and openness have been related to broad searching of diverse sources, and openness and conscientiousness have been positively associated with deep searching (Heinström 2006), extraversion, openness, and conscientiousness should be positively associated with those plans' complexity. Further, self-monitoring and cognitive complexity-linked to plan complexity in previous research (Berger 1997: 42-46)-will probably also be positively associated with information-seeking-plan complexity.
Once activated, the information-seeking plan should spur the performance of behaviors represented in it (Greene 1984). Accordingly, citizens' informal political talk with friends and family during a major political event should feature information-seeking behaviors: questions about political choices and members' decision-relevant opinions and values. 25 Further, because more complex plans describe greater numbers of behaviors (Berger 1997: 28-29), plan complexity should be positively associated with the frequency of performing communicative behaviors. Consequently, traits positively associated with the complexity of information-seeking plans-extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity-should be positively associated with the frequency of using information-seeking behaviors. Neuroticism-predicted to be negatively associated with information-seeking plan complexityshould be negatively associated with the frequency of those behaviors. Moreover, because, on this account, plan complexity is expected to channel some influence of traits on the performance of information-seeking behaviors, those associations should be partially mediated by plan complexity.
On obtaining the needed information, a second message-production phase begins ( Table 1). That information should enable the citizen to feel they have made sense of the political choice (Richards 2018) and to make the choice, which will probably increase their decision-making confidence-a kind of internal political efficacy (Zhang & Chang 2014). 26 The citizen should then feel more capable of helping others make political choices and of persuading others to choose in particular ways. Further, the citizen should perceive undecided family or friends or those who made a different political choice as opportunities to exert influence (Dillard & Solomon 2000: 170). That perception should cause cognitive rules to activate goals of providing advice and changing others' opinions (Kellermann 2004), which should join decision-making and informationseeking goals in defining citizens' informal political discussions in the Informal Major-Political-Event Context: H2(a). Goals of making a political decision, obtaining information, providing guidance/advice, and changing others' opinions will define citizens' interactions to a greater extent than goals of maintaining relationships, reputation, or self-image, passing the time, self-development, or stimulating interesting discussion.
The advice-giving and opinion-changing goals should activate a justification plan (Berger 1997: 24). Incorporating elicited information about close-socialnetwork members' political views and values, this plan should contain representations of behaviors involving justifying the citizen's political decision and influencing network members' decisions on grounds acceptable to those members. Thus in the Informal Major-Political-Event Context, in informal political talk with family and friends, H2(b). Citizens will plan to use (i) informationseeking and (ii) reason-giving behaviors more often than in the Informal Ordinary Context.
Further, in the Informal Major-Political-Event Context, self-monitoring and cognitive complexity-related in earlier research to plan complexity (Berger 1997: 42-46)will probably be positively associated with the justification plan's complexity.
Plan activation should trigger performance of reasongiving behaviors represented in the plan (Greene 1984). Consequently, in this context, the citizen's informal political talk with close-social-network members will probably feature reason-giving (Neblo 2015): advising or persuading members about their political decisions and justifying the citizen's decision on grounds acceptable to those members. Thus during such talk in this context, H2(c). Citizens will use (i) information-seeking and (ii) reason-giving behaviors more often than in the Informal Ordinary Context.
Because plan complexity should be positively associated with frequency of performing communicative behaviors, in this context, self-monitoring and cognitive complexity (Berger 1997: 42-46) will probably be positively related to frequency of using reason-giving behaviors, and justification-plan complexity should partially mediate those relationships.
Because most major political events impose decisional deadlines, citizens' informal discussions with family and friends during such events should occur more frequently than during the Informal Ordinary Context, and during major elections should occur regularly because those elections generally recur at consistent intervals. Therefore information-seeking and justification plans activated in informal political discussions during major political events should be stored in memory, repeatedly re-used, and elaborated over lifetimes (Berger 1997: 26, 47). 27 After several major political events, even citizens without formal debate training should possess fairly well-elaborated plans specifying behaviors for justifying political decisions to close-social-network members and should have gained experience in reason-giving about political decisions in discussions with those members (Warner & Colaner 2016).

Formal Contexts
The GSJ model specifies two contexts involving organized deliberation with strangers: outside of major political events (Formal Ordinary Context) and during such events (Formal Major-Political-Event Context) (collectively, Formal Contexts). 28 Formal Contexts include deliberative meetings with legislators (Neblo et al. 2018) and other mini-publics (Smith 2009). Some formal deliberations require voting or consensus (Sprain & Gastil 2013); whereas, others do not (no-vote/no-consensus processes) (Fishkin 2009).
The GSJ model predicts the Informal Major-Political-Event Context and Formal Contexts are sufficiently congruent that similar goals will be activated in both contexts, which should trigger the same plans, which should produce similar political-communicative behaviors, as follows. Both contexts apply pressure to make a political or policy decision on citizens who lack sufficient information to make the decision and to justify the decision to group members. In both contexts, citizens consequently experience needs for information required for making and justifying the decision. In both contexts, citizens are likely to perceive the decision-making task and information needs as opportunities and obstacles, which should activate decision-making and information-seeking goals. In both contexts, acquiring that information and making the decision should boost citizens' sensemaking and self-efficacy and enable citizens to formulate justifications for their decision. In both contexts, citizens should perceive group members as affording opportunities to exert influence, which should activate goals concerning advising and persuading. We now explain in detail the GSJ model in Formal Contexts.

Formal Ordinary Context
In the Formal Ordinary Context, deliberative processes involve making decisions, whether group choices (Gastil & Knobloch 2020) or individual judgments (Fishkin 2009). Citizens typically lack knowledge relevant to those decisions due to the issues' ad-hoc nature (e.g., unexpected crimes) and citizens' paucity of political knowledge (Delli Carpini & Keeter 1996). Moreover, citizens often lack awareness of the process's required standards and other unfamiliar participants' views and values concerning the focal issue. Under these conditions, goal activation should proceed in two phases.
Information-seeking in the Formal Ordinary Context begins with the formal process's decision-making task, which makes decision-making salient for citizens. Decision-making presents opportunities to influence policy and to satisfy the formal process's administrators' and other participants' desires for a decision (Dillard & Solomon 2000). That perception should cause cognitive rules to retrieve the decision-making goal as a primary goal (Wilson 1990) ( Table 2).
Activating that goal should raise citizens' awareness of needs for information about the process's issue and evaluative criteria and other participants' issue-relevant views and values. Those needs should trigger emotions of confusion, anxiety, and disorientation (Kuhlthau 1991). Thus emotion should be integral to citizens' formal deliberations (Neblo 2015). To quell these feelings, citizens should seek information to fulfill those needs. Perceiving those needs as obstacles to decision making (Dillard & Solomon 2000) should cause cognitive rules to activate the information-seeking goal as a primary goal (Wilson 1990).
Moreover, in no-vote/no-consensus processes, the absence of social pressure to make a public political decision should reduce the extent to which decisionmaking defines participants' interactions. The citizen should perceive that attenuated decision-making imperative as an opportunity to pursue personal objectives (Dillard & Solomon 2000), which should cause cognitive rules to activate the self-development goal (Wilson 1990). Thus self-development (Conover et al. 2002) will likely join decision-making and information-seeking goals in defining those interactions in no-vote/no-consensus processes in the Formal Ordinary Context: 29 Note. An asterisk indicates a primary goal that is likely to define participants' interactions in formal deliberative processes that do not require participants to vote or reach consensus on a decision; in other types of formal deliberative processes, that goal is likely to serve as a secondary goal. The goal called Obtain information includes the goal of gathering knowledge. The personality trait of neuroticism is expected to be negatively associated, and the personality traits of extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity are expected to be positively associated, with the complexity of information-seeking plans and with the performance of information-seeking behaviors; information-seeking-plan complexity is expected to mediate partially associations between traits and performance. The personality traits of self-monitoring and cognitive complexity are expected to be positively associated with the complexity of justification plans and with the performance of reason-giving behaviors; justification-plan complexity is expected to mediate partially associations between traits and performance. See text for details.
Richards and Neblo: Active Is as Active Does 8 H3. The self-development goal will define citizens' interactions to a greater extent than goals of maintaining relationships, reputation, or self-image; passing the time; or stimulating interesting discussions.
Because during past major political events the citizen likely elaborated and stored in memory such an informationseeking plan, in this context the information-seeking goal should trigger the retrieval from memory of that plan, which should then be adapted for use in the formal process (Berger 1997: 26). 30 Due to ' cognitive processing limitations,' citizens will likely reuse a pre-existing plan even if it only roughly fits the current setting (Berger 1997: 26). 31 Adaptations citizens will likely make to that plan (Berger 1997) include adding representations of behaviors involving inquiring about information concerning the process's focal issue and required evaluative criteria. As in the Informal Major-Political-Event Context, in the Formal Ordinary Context neuroticism should be negatively associated and extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity positively associated with information-seeking-plan complexity (Berger 1997;Heinström 2006).
The retrieved and modified plan should trigger the performance of information-seeking behaviors represented in the plan (Greene 1984), and the citizen's interactions in the Formal Ordinary Context should feature information-seeking behaviors: questioning of witnesses about the issue, administrators about evaluative criteria, and other participants regarding their issuerelevant opinions and values.
As in the Informal Major-Political-Event Context, in the Formal Ordinary Context, citizens' extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity (Berger 1997: 42-46;Heinström 2006) should be positively-and citizens' neuroticism negativelyassociated with the use of information-seeking behaviors, and information-seeking plan complexity should partially mediate those associations.
Acquiring the desired information should trigger the second, justification, phase ( Table 2). That information should help the citizen make sense of the decision, take the decision, and experience greater confidence in decision-making (Richards 2018). Moreover, the acquired information concerning evaluative criteria and other participants' views and values should help the citizen fashion justifications likely to satisfy those criteria and to be acceptable to other participants. The citizen then should feel more confident in advising or persuading other participants about their decisions and in justifying his/her decision to those participants.
As in the Informal Major-Political-Event Context, the citizen, in the Formal Ordinary Context, is likely to perceive other participants who remain undecided or whose decision differs from the citizen's as opportunities for exerting influence (Dillard & Solomon 2000). That perception should lead cognitive rules to activate goals of providing advice and changing others' opinions (Kellermann 2004). Consequently, those goals should accompany the decision-making and information-seeking goals-as well as, in no-vote/no-consensus processes, the self-development goal-in defining the citizen's interactions in the Formal Ordinary Context.
Then advice-giving and opinion-changing goals should activate the pre-existing justification plan. Developed over repeated major political events, that plan should be retrieved from memory and adapted for use in the Formal Ordinary Context, even if that plan only approximately suits that setting (Berger 1997: 24-26). That plan should spur performance of behaviors (Greene 1984) involving justifying the citizen's decision and persuasive claims with reasons designed to satisfy the process's required criteria and to be acceptable to other participants. Because the GSJ model predicts citizens' political communication in the Formal Major-Political-Event Context will track that in the Formal Ordinary Context, with a few exceptions described below, we hypothesize, in political-communicative interactions in Formal Contexts H4(a). Goals of making a political/policy decision, obtaining information, providing guidance/advice, and changing others' opinions will define citizens' interactions to a greater extent than goals of maintaining relationships, reputation, or self-image; passing the time; stimulating interesting discussion; or-except in the Formal Ordinary Context in processes that do not require participants to vote or reach consensus-self-development. H4(b). Citizens will plan to use (i) informationseeking and (ii) reason-giving behaviors more often than in informal political-communicative interactions with close-social-network members in the Informal Ordinary Context, and as often as in such interactions with those members in the Informal Major-Political-Event Context. H4(c). Citizens will use (i) information-seeking and (ii) reason-giving behaviors more often than in informal political-communicative interactions with close-social-network members in the Informal Ordinary Context, and as often as in such interactions with those members in the Informal Major-Political-Event Context.
Moreover, because priming is conventionally used to investigate plans (Berger 1997: 47-51), the GSJ model predicts, in Formal Contexts H4(d). Citizens-primed to recall informal political-communicative interactions with close-socialnetwork members that occurred in the Informal Major-Political-Event Context-will use more (i) information-seeking and (ii) reason-giving behaviors in communicative interactions than citizens not so primed.
As in the Informal Major-Political-Event Context, in Formal Contexts, citizens' self-monitoring and cognitive complexity (Berger 1997: 42-46) should be positively associated with frequency of using reason-giving behaviors, and plan-complexity should partially mediate those associations. Therefore-also taking into consideration the account above of relationships between personal traits and information-seeking-plan complexity-in citizens' political-communicative interactions in the Informal Major-Political-Event Context and Formal Contexts H5. Citizens' extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity will be positively associated, and citizens' neuroticism will be negatively associated, with (i) the complexity of information-seeking plans and (ii) the frequency of performing information-seeking behaviors. (iii) The complexity of informationseeking plans will partially mediate associations between traits and behaviors. H6. Citizens' self-monitoring and cognitive complexity will be positively associated with (i) the complexity of reason-giving plans and (ii) the frequency of performing reason-giving behaviors. (iii) The complexity of reason-giving plans will partially mediate associations between traits and behaviors.

Formal Major-Political-Event Context
Citizens' political-communicative interactions in the Formal Major-Political-Event Context are expected to differ somewhat from the Formal Ordinary Context. During a major political event, in no-vote/no-consensus processes whose focal issue concerns the event, citizens are likely to perceive decision-making concerning that issue as salient and as an opportunity to influence policy and to placate those exerting pressure to decide. That perception should lead cognitive rules to activate the decision-making goal to a greater degree than in such processes during the Formal Ordinary Context. Further, in the Formal Major-Political-Event Context, decision-making's heightened salience should cause citizens to perceive fewer opportunities to pursue personal objectives, leading cognitive rules to attenuate activation of the self-development goal. Moreover, the major political decision's salience should increase citizens' perceptions of opportunities to affect that decision by influencing others' choices on the focal issue. Accordingly (Table 3), H7. Goals of making a political/policy decision, providing guidance, and changing others' opinions will define citizens' interactions to a greater extent than in such processes in the Formal Ordinary Context.

Preliminary Evidence
We now assess the GSJ model's congruence with results from previous and new research.

Evidence: Informal Ordinary Context
Regarding goals, studies showed social objectives, such as passing time with others, and personal objectives, such as self-development, defined citizens' informal political talk outside of major political events (Conover et al. 2002: 51-53;Gil de Zúñiga et al. 2016: 539-540). These findings generally support Hypothesis 1(a)'s predictions about social and personal goals as primary goals. Yet results concerning deliberative objectives, like persuasion, were mixed. 32 With respect to behaviors, Walsh (2004) found most political discussions arose during everyday conversation and featured more identity-expressions than reasongiving. The former result supports the GSJ model's account of the Informal Ordinary Context and the latter Hypothesis 1(c)'s reason-giving prediction. Moreover, that many U.S. citizens do not know their familymembers' partisan identifications (Bennett et al. 1995: 286)-suggesting a paucity of reason-giving behaviors in the Informal Ordinary Context that might reveal family-members' political views-supports that same prediction.

Evidence: Informal Major-Political-Event Context
Regarding goals, research indicated decision-making and information-seeking defined citizens' informal political talk during major elections (Bennett 1994: 12;33 Eveland et al. 201133 Eveland et al. : 1091Morey & Yamamoto 2020: 89-91). Yet regarding persuasion and advice-giving as primary goals and social and personal goals as secondary goals, evidence was inconsistent (compare Eveland et al. 2011: 1091with Morey & Yamamoto 2020. Those results support Hypothesis 2(a)'s predictions about information-seeking and decision-making as primary goals. Yet findings are mixed regarding predictions concerning persuasion, advice-giving, and social and personal goals.
In addition, studies indicate many citizens, as noted above, do not know their family members' political affiliations, political talk with family and friends is more frequent during major elections-especially among citizens less likely to be politically knowledgeable-and political-discussion frequency is positively associated with political-decision salience, willingness to vote, and voter turnout (Bennett et al. 1995: 286, 289-292;Binder et al. 2009;Kenney 1993: 232;Kim et al. 1999: 378). Moreover, research shows citizens' political-conversation frequency is positively associated with argument quality, and willingness to argue about politics is positively related to perceiving oneself as in the majority in one's close social network (Kim et al. 1999: 375, 377). This evidence supports Hypothesis 2(a)'s predictions concerning decisionmaking and information-seeking as primary goals and the GSJ model's account of an information-seeking phase involving learning close-social-network members' views, followed by a justification phase.
Regarding behaviors, studies show friends' furnishing reasons to justify major-election votes (Cramer 2016: chapter 7), more citizens report political proselytizing in major elections than in other elections (Bennett et al. 1995: 281), and, as noted above, citizens' political-talk frequency is positively associated with argument quality (Kim et al. 1999). Collectively, that evidence accords with the reason-giving prediction in Hypothesis 2(c). Note. An asterisk indicates goals which, in formal deliberative processes that do not require participants to vote or reach consensus on a decision and whose topics are related to the major political event, the GSJ model predicts will be rated as defining communicative interactions to a greater extent than in the Formal Ordinary Context. The goal called Obtain information includes the goal of gathering knowledge. The personality trait of neuroticism is expected to be negatively associated, and the personality traits of extraversion, openness, conscientiousness, self-monitoring, and cognitive complexity are expected to be positively associated, with the complexity of information-seeking plans and with the performance of information-seeking behaviors; information-seekingplan complexity is expected to mediate partially associations between traits and performance. The personality traits of self-monitoring and cognitive complexity are expected to be positively associated with the complexity of justification plans and with the performance of reason-giving behaviors; justification-plan complexity is expected to mediate partially associations between traits and performance. See text for details. Jacobs et al. (2009: 73) found social motivations generally served as secondary goals for citizens' participation in formal deliberations, consistent with Hypothesis 4(a)'s prediction that social goals would be secondary goals.

Evidence: Formal Contexts
Regarding information-seeking and sense-making, previous research showed citizens gained issue-knowledge and political efficacy during formal deliberations (Esterling et al. 2011: 498;Gastil & Dillard 1999: 186-187;Price & Cappella 2002: 317-319). Further, we analyzed original survey data from the Oregon Citizen Assembly on COVID-19 Recovery. 34 Results showed participants' sense-making about issues and policies and confidence in making recommendations increased significantly throughout the information-gathering phase, and sensemaking and confidence were positively and significantly correlated. 35 Collectively, these findings support the GSJ model's account of how information-seeking and discussion foster increased sense-making and efficacy in formal deliberations.
With respect to behaviors, studies report substantial reason-giving in citizens' formal deliberations (Gerber et al. 2018(Gerber et al. : 1106Polletta & Lee 2006: 710;Stromer-Galley 2007: 15-16), but findings about frequency of information-seeking behaviors are mixed (compare Gerber et al. 2018: 1106with Stromer-Galley 2007. These results accord with Hypothesis 4(c)'s prediction about reason-giving but furnish inconsistent support for the information-seeking prediction. Further, these studies did not compare results with the Informal Major-Political-Event Context.
Regarding relationships between informal and formal political discussion, research showed jurors reused deliberative norms and practices developed outside the courts (Sprain & Gastil 2013;Sunwolf & Seibold 1998). Other scholars, finding informal political discussion with close-social-network members was positively associated with subsequent participation in formal deliberations, characterized those results as evidence that informal political talk with family and friends trains citizens for formal deliberation (Conover et al. 2002: 37-38;Schmitt-Beck & Grill 2020: 12, 16). 36 These findings broadly accord with, while not furnishing direct evidence supporting, Hypothesis 4(d).
Finally, no evidence was found relevant to Hypotheses 3 or 5 through 9.
Considered together, the foregoing evidence indicates the plausibility of the GSJ model's core principles.

Discussion
The goals-sense-making-justification (GSJ) model of citizens' political communication draws on messageproduction (Dillard 2015) and sense-making (Dervin 1989) theories to explain why such communication exhibits deliberative characteristics in some contexts but not others. The model predicts different contexts of political discussion cause different goals to define citizens' political-communicative interactions, which activate different plans, which trigger different politicalcommunicative behaviors.
In decision-making contexts, the model predicts goals defining citizens' interactions and plans serving those goals lead citizens to engage in information-seeking and in reason-giving characteristic of deliberation. Yet in nondecisional settings, such as informal political talk outside of major political events, the model holds that citizens' goals and related plans encourage the exchange of messages like tokens during small talk and make less likely the kind of reason-giving that distinguishes deliberation. Thus the GSJ model reconciles claims of theorists arguing citizens' political talk is non-deliberative (e.g., Eveland et al. 2011) with those holding such talk should be deemed deliberative (e.g., Neblo 2015). 37 Moreover, the model illustrates multiple goals' shaping citizens' political communication (Blumler 2015: 433-434).
Evidence supports many of the model's principles, including phases of information-seeking and justification in informal political discussion with family and friends during major political events, and that informationseeking fosters sense-making and empowerment during such events and formal deliberations. Findings support hypotheses concerning social and personal goals as primary goals and infrequent reason-giving in informal political talk in close social networks in the Informal Ordinary Context, and regarding the primacy of decisionmaking and information-seeking goals in such talk during major political events. Evidence also supports predictions about the frequency of reason-giving in such talk during such events and formal deliberations. Further, research broadly accords with the model's characterization of informal political talk as training citizens for formal deliberations.
Yet evidence regarding some aspects of the model is mixed or absent. Results are inconsistent concerning the roles of persuasive and advice-giving goals in informal political discussion during and beyond major political events. No evidence has been located concerning the selfdevelopment goal's role during formal deliberations, or about whether participants prioritize decision-making and persuasive goals during formal processes to a greater extent during major political events than beyond such events. Nor has evidence been identified concerning plans in citizens' political communication, and evidence of the frequency of information-seeking behaviors is mixed or absent. Moreover, no evidence has been found concerning personality traits' influence on citizens' political-communicative plans or behaviors.
Features of the GSJ model about which there is no or inconsistent evidence should be foci of future research. Consistent goals (Morey & Yamamoto 2020) should be analyzed in informal settings-among family and friendsand formal contexts regarding those goals' primary or secondary roles and associations with plans and behaviors. In addition, research should explore personality traits' influence on plan complexity and frequency of information-seeking and reason-giving behaviors (Berger 1997;Heinström 2006).
Theoretical limitations also characterize this account. First, whereas this account defines deliberation as reasongiving, future theorizing should extend the model to incorporate other indicia of deliberation (e.g., Habermas 1996). Although this account details how emotions influence information-seeking, future GSJ-model research should encompass emotions in reason-giving (Neblo 2020). Further, this account of the Informal Major-Political-Event Context covers only communication with family and friends. Future study of that context should address political-canvassing interactions (Kaid 1977). 38 Though this account does not distinguish in-person from online political talk, nonetheless, because deliberations in those modalities may differ (e.g., Stromer-Galley et al. 2015), adapting the model to account for those differences should be explored in future research. Moreover, whereas this account explains how uninformed citizens learn to decipher partisan cues that fuel political polarization, future theorizing involving the GSJ model should address polarization's effects on citizens' deliberations (Cramer 2016). 39

Conclusion
Theorizing about citizens' political communication has been stymied by disagreement over whether deliberative or non-deliberative theory best characterizes such communication. The goals-sense-making-justification model presented in this article helps to overcome that impasse. This article demonstrates how accounting for cognitive phenomena that generate political communication can help advance theorizing concerning citizens' deliberations.

Notes
1 'Political communication' means symbolic exchanges about public policy or public affairs. 2 'Citizens' denotes members of a political community regardless of naturalization. 3 'Public policy' means collective action to address matters of community interest. 4 'Politics' denotes the use of power in relation to institutions. 5 'Message production' means the goal-oriented process of creating symbols and behaviors aimed at expressing intentions and emotions to others (e.g., Dillard 2015). 6 The goals-plans-action (GPA) model arose through research analyzing face-to-face and online interactions (see Dillard 2004Dillard , 2015, while individual-level sensemaking theory sought to account for informationseeking using human, print, and online sources (Dervin 1989;Kuhlthau 1991). 7 We return to this topic in the Discussion section. 8 We consider sense-making at the individual (Dervin 1989) rather than the organizational level because most principles of our model concern informal communication during everyday life, and most formal citizen-deliberation processes use 'zero-history' groups, which lack the group awareness, norms, and experience (Gastil 2010) that enable group-level or organizational-level sense-making. 9 'Context' is a synonym for 'setting' (Merriam-Webster, n.d.). 10 For empirical evidence of the causal influence of contexts on goals, see Wilson (1990); Hample (2016). 11 We thank an anonymous reviewer for suggesting the inclusion of the self-development goal (Conover et al. 2002). 12 For empirical evidence of the causal relationship between goals and plans, see Berger (1997); Dillard (2015). 13 For empirical evidence of the causal relationship between plans and communicative behaviors, see Berger (1997). 14 For the theoretical explanation for reuse of plans, see note 27 below. 15 In this article, major elections are always discussed as examples of the broader category of major political events defined below. 16 In this article, 'should' is 'used in an auxiliary function to express what is probable or expected' (Merriam-Webster, n.d.). That is, in this article, 'should' is used in a spirit of epistemic modesty to convey what is likely but uncertain. 17 Because theory suggests that this explanation is not limited to any particular context (see Berger 1997: 24), this explanation is not repeated below. 18 Because theory does not indicate that this explanation is limited to any particular context (Greene 1984), this explanation is not repeated below. 19 To be sure, the GSJ model does not preclude the possibility that well-structured conversations among individuals may occur at any time. Nonetheless, as stated earlier in this article, the GSJ model is designed to explain commonly observed phenomena in most citizens' everyday political communication, which, according to the published empirical literature, is generally woven into small talk and generally does not feature the reason-giving that is characteristic of deliberation (e.g., Eveland et al. 2011;Goldsmith & Baxter 1996;Walsh 2004). 20 The GSJ model centers on informal political talk among family and friends because empirical politicalcommunication research consistently finds that ordinary people in democracies talk most often about politics with members of their close social networks (e.g., Eveland et al. 2011Eveland et al. : 1090. Further, during majorelection campaigns, relatively few U.S. residents engage in political canvassing, which would enable them to engage in political discussion with those outside of their close personal networks; for example, in 2018, just 16% of U.S. respondents reported that they had 'worked or volunteered for a political campaign in the past five years' (Pew Research Center 2018: 101). To be sure, online networks offer citizens opportunities to engage in informal political discussion with individuals beyond family and friends. Online political discussion, which is not directly addressed in this article, is addressed in the Discussion section below. 21 Major elections-elections in which most citizens experience social pressure to vote-include, in presidential systems, presidential elections, and, in parliamentary systems, general elections, though they sometimes include other elections, such as the 1866 U.S. midterm election (Burnham 1999). Social pressure to decide, during major elections, is evident in greater media coverage of politics (e.g., Bennett 1994: 3, 11), voters' greater concern about election outcomes (Bennett 1994: 12), and higher voter turnout (Stanley & Niemi 2015: 4-6) than in other elections. Relatedly, during major elections, citizens talk more about politics with family and friends than in other periods (Binder et al. 2009: 323-324;Bennett et al. 1995: 282, 289). 22 In democracies with compulsory voting, laws also exert such pressure (Fishkin 2009). 23 The GSJ model is consistent with theory and evidence concerning how political polarization influences citizens' political communication and decision making. Accounts of many citizens' lack of basic political information (e.g., Delli Carpini & Keeter 1996) indicate that many citizens, at the outset of a major political event, are incapable of interpreting partisan cues that enable polarized thinking and communicating, because those citizens are unaware of major parties' ideological and policy positions. The GSJ model explains why and how citizens gain additional, basic political information that will, among other things, enable those citizens to decipher partisan cues and thus to engage in polarized political thinking and communicating, as well as non-polarized political thinking and communicating, during a major political event. That is, the GSJ model offers one explanation for how those cues become intelligible and useful to such citizens. 24 The GSJ model does not predict that citizens generally will seek diverse information, although the model acknowledges previous research showing personality traits of extraversion and openness have been positively associated with broad searching of diverse sources (Heinström 2006). Rather, the GSJ model predicts-consistent with findings from sense-making research-that citizens generally will seek information that will satisfy their information needs (Dervin & Frenette 2001;Kuhlthau 1991). Further, for some citizens-especially first-time or non-habitual voterssuch anxiety may cause non-voting or withdrawal from politics (Weber 2013). This pattern suggests weak voting norms in these citizens' close social networks. The GSJ model concerns the majority of citizens who operate under stronger voting norms and who tend to respond to this anxiety by seeking information. 25 To be sure, citizens' cognitive or perception-consistency biases are likely to influence citizens' informationseeking behavior and citizens' cognitive processing of collected information (Sunstein 2018). Such biases may reduce the quality of the information collected, but the GSJ model does not assume that such biases will prevent citizens from gathering the information required to satisfy their information needs, make political decisions, or develop justifications for those decisions that are acceptable to members of the citizens' close-social network. The risk that citizens' deliberations and political decisions will be influenced by inaccurate information and cognitive biases and heuristics is inherent to open, democratic societies (see, e.g., Gastil 2008). 26 For empirical evidence of positive associations between sense-making and efficacy, see Kuhlthau (1991) and note 35 below. 27 As explained above, such reuse is likely because the mind, to conserve limited cognitive resources, after goals are activated initially searches memory for an existing plan depicting behaviors serving activated goals and then modifies that plan for the current context (Berger 1997: 26, 47). 28 We acknowledge an anonymous reviewer who strongly urged this bifurcation of the formal context. 29 We acknowledge an anonymous reviewer who made these points. 30 This prediction concerning reuse of communication structures and practices, originally developed in one context, in a different context accords with prior research. According to Warner and Colaner (2016: 207), citizens' informal political deliberation with family and friends equips citizens with 'resources they need to be effective deliberators' in formal settings. Structuration scholars predicted group participants would reuse, in their groups, argument structures and practices-'argument systems' (Gastil 2010: 73-74)developed outside of their groups and furnished evidence of jurors' reuse of deliberative rules and practices developed outside of courts (Sprain & Gastil 2004;Sunwolf & Seibold 1998). Further, analysts of citizens' political communication characterized their findings as evidence that citizens' informal political talk with family and friends trains citizens for formal deliberation (Conover et al. 2002: 16;Schmitt-Beck & Grill 2020: 37-38). 31 Berger (1997: 26) explains that 'because of … cognitive processing limitations…, people probably overgeneralize similarity when the fit between canned [i.e., already-developed] plans and the current situation is relatively close. Individuals may be prone to overlook subtle differences between their canned plans and the current exigencies entailed by the social context.' 32 Gil de Zúñiga et al. (2016: 539) found that ' civic' goals, including 'persuad[ing] others,' were positively associated with informal political-discussion frequency, whereas Conover et al. (2002) found that decision making and persuading others were infrequently mentioned as informal-political-discussion goals. 33 Bennett (1994:12) found substantially more citizens 'personally cared which party won the election'-a gauge of decision-making salience-during a presidential than an off-year election. 34 The Assembly (https://healthydemocracy.org/covid/), which featured information-gathering and deliberative phases, was held online on Zoom, July 9-August 20, 2020, in seven, two-hour sessions. Participants were a stratified random sample of 36 adult residents of the U.S. state of Oregon charged with identifying key issues related to recovery from the COVID-19 pandemic and making policy recommendations about those issues to the Oregon state legislature. Participants completed a survey after each Assembly session. 35 Here are detailed survey results concerning sensemaking and political efficacy. The degree to which participants could make sense of the issue of COVID-19 recovery was measured with a five-point Likert-type scale ( ; paired-samples t-test: t (26) = 2.12, p = 0.041, twotailed. We also estimated the zero-order correlation between participants' sense-making about the COVID-19 issue and participants' confidence in their ability to make recommendations to their state legislature about COVID-19 recovery after the Week 6 Assembly Session: r = 0.65, p < 0.001, two-tailed, N = 28. In addition, we estimated the zero-order correlation between participants' sensemaking about policy options for addressing COVID-19 recovery and participants' confidence in their ability to make recommendations to their state legislature about COVID-19 recovery after the Week 6 Assembly Session: r = 0.72, p < 0.001, two-tailed, N = 28. 36 Conover et al. (2002: 37-38) Neblo (2015) also argues that political communication that is not deliberative on its face (in the sense of reason-giving) may yet contribute to the larger deliberative system. 38 We thank an anonymous reviewer for this point. 39 We thank an anonymous reviewer for this point as well.