The Discourses of Welfare and Welfare Reform

 

John W. Mohr

 

The study of social welfare has undergone a shift over the last 20 years from a strongly realist to a decidedly constructionist orientation. The move is largely the result of the impact of feminist scholars such as Nancy Fraser and Linda Gordon whose attention to the construction of gender categories called into question key analytic assumptions of earlier research agendas. The cultural turn that took place in this arena depended on the analysis of discourse. In this chapter I explain how the concept of discourse came into and subsequently transformed the sociological study of welfare institutions. I then highlight two key features of institutional discourse that I believe need to be taken into account in future research — that it is: (1) organized within semiotic systems, and (2) constructed through mutually constitutive dimensional orders. I will develop these arguments by highlighting some findings from my own research on the history of the American welfare state.

Sociology, Social Welfare, and Social Realism

 

            There is a longstanding connection between sociological research and the field of social welfare. These linkages were especially evident during the Progressive Era when academic departments of sociology were being founded in American universities at the same time that the profession of social work was being established and the social welfare sector was undergoing a period of intense rationalization. This conjuncture was defined by a style of pragmatic realism; sociologists saw the field of social welfare as a place in which their theories and professional expertise (on the nature of inequality, the causes and consequences of social problems, the rules of social organization, etc.) would find practical application (Furner, 1975; Ross, 1991).

            After the second world war, American sociology underwent a paradigm shift manifested as a refusal of a common theoretical framework, in particular, a rejection of grand theory in favor of something that Robert Merton (1957) called theories of the middle range. The emergent paradigm built upon a common epistemology and a methodological imperative that embraced the systematic interrogation of observational data (both quantitative data as pioneered by Lazarsfeld and his colleagues but also in the qualitative analysis of social phenomena as expressed by the phenomenological turn of scholars such as Harold Garfinkel, Aaron Cicourel, and Harvey Sacks). What followed was an enormously productive period of growth in American sociology and a shift of the core sub-fields of the discipline toward the embrace of formal modeling methodologies. These advances came at a price. The work of science is difficult. Small variations in measurement have to be noticed , managed statistically and incorporated into an explanatory frame. This demands specialization and an intense narrowing of the field of vision. To maintain its forward inertia science must be well articulated with a theoretical model. Thus as methods became more carefully refined, core theories began to shrink. Specialists in theory responded by devoting much of their energy to philosophical studies of method, epistemological battles over the proper domain of science, and meta-level theories of action and practice.

            In the literature on social welfare institutions, the new empirically oriented research paradigm first found expression in the work of scholars like Harold Wilensky and Charles Lebeaux whose book Industrial Society and Social Welfare (1958) became a classic. Standing at the beginning of what has become a long and fruitful tradition of research, Wilensky and Lebeaux linked together formal methods for summarizing broad scale indicators of social activity with a modified structural functionalism in which welfare was operationally defined as “those formally organized and socially sponsored institutions, agencies and programs, exclusive of the family and private enterprise, which function to maintain or improve the economic conditions, health or interpersonal competence of some parts or all of a population,” (1958, p. 17). They argued that changes from one level of welfare provision to another were propelled by the forces of modernization — a concept they referred to as the “logic of industrialism”.

Other perspectives soon emerged to counterbalance this explanatory frame. Neo-marxists (O'Connor, 1973), state capacity theorists (Skocpol and Ikenberry, 1983), power resource theorists (Esping-Anderson and Korpi, 1984) and institutionalists (Thomas, Meyer, Ramirez, and Boli, 1987) provided alternative explanations. All of these projects (lets call them formal comparativists) share the same underlying form. Variations in the level of provision of social benefits constitute the dependent variable. Theoretical disputes revolve around differences in independent variables (features of the society that are hypothesized to explain this variation). The “logic of industrialism” perspective highlights the importance of economic and technological development as the root cause for the growth of welfare states. State autonomy theorists emphasize the character and capacities of state institutions, as well as the expertise and vested interests of state bureaucrats. Institutionalists measure features of the global social order and a particular country's relationship to that order. These explanation are framed within conventions of understanding that are dually articulated as theoretical constructs and methodological practices embedded within a particular scientific habitus (Mohr, forthcoming). Concrete, well-metered features of the demography, the economy, the polity, the political economy or position within the world order thus come to be defined as the primary (e.g., genuine) causal factors leading to the growth of welfare activities.

“Objective” measurability is also a criterion for what comes to be conceptualized as a viable dependent variable. Initially, variation among welfare states was measured by overall expenditures in social programs. Over time these measures improved. First, outcomes were broken into types and taxonomies and data collection came to focus on the presence and absence of particular categories of social provision or, more recently, on the types of eligibility criteria that citizens can use in making claims against the state (Esping-Anderson, 1990). While both the measures and the models have become more sophisticated, so too have the theories improved. And yet, like so much of the social science that emerged during this period, this research often leaves one feeling lost in a sea of details, holding onto explanatory narratives that feel overly neat, analytic and abstract (see Mohr, 1998 for a critique). We can notice this now thanks to the cultural turn that has taken place, largely at the behest of feminist scholars.

The Cultural Turn

 

Feminist scholarship begins from the perspective that what matters is not the objective quality of sex, but the cultural system of meanings embodied in gender relations. Thus culture shapes and conditions social distinctions that come to be treated as objective. When feminist scholars took up the study of the welfare state they brought this sensibility to bear and quickly began to emphasize how the social categories that underlie most welfare systems—especially gendered categories such as "widow," "mother," "unwed mother," and the like—are symbolic constructs that contain within them ideologically coded assumptions about gender roles, the concept of a "family wage," the proper separation of public from private spheres as well as many other morally charged cultural prejudices. A classic example is how single mothers in the United States (in their roles as “beneficiaries” of federal support programs such as the Food Stamp or AFDC programs) have traditionally been held to a type of moral policing that was not imposed on the beneficiaries of masculinized relief programs (such as unemployment assistance or retirement insurance). Whether one looks at the food stamp program in which relief applicants are given content-coded stamp books rather than cash as a way to control spending habits or AFDC type support programs in which a relief recipient's sexual life is considered an appropriate object of scrutiny, feminized social welfare programs tend to view women as being in need of close moral supervision. In Nancy Fraser’s terms, “welfare practices construct women and women's needs according to certain specific—and in principle, contestable—interpretations, even as they lend those interpretations an aura of facticity that discourages contestation” (1989, p. 146).

            The idea that social welfare systems are charged with interpretative ambiguities and a rich moral discourse is not new. Historians have long emphasized how the contestation over meanings—distinctions between the worthy and the unworthy poor, for example—are fundamental features of welfare institutions. But this type of interpretative sensitivity was considerably dulled by the formal comparativists whose embrace of quantification encouraged an uncritical and unrefined acceptance of received categories. Thus, what the feminists inaugurated was a substantive change, a cultural turn that shoved the sociological study of welfare institutions away from positivist duels over which regression line best fit what data towards a scholarly orientation in which welfare programs are investigated as “institutionalized patterns of interpretation” (Fraser, 1989). Feminist scholars highlight the complex ways in which identities are differentiated, eligibilities are given alternative moral weightings, needs are socially constructed and benefit programs are far more complex than measures of levels of GNP could express. Indeed, one especially important contribution of feminist scholars was to remind us of how welfare demeans women at the same time that it supports them, and thus to reveal the ways in which welfare programs can be intrusive, coercive or controlling as well as respectful, enabling or liberating.

One way to read this is to say that feminists raised the bar on the expectations we have for providing what Geertz (1973) would call a thick rather than a thin description of welfare policies. They argued that without an interpretative analysis we are constrained in our ability to understand and, thus, to effectively compare (formally or otherwise) the social policies that are enacted at different times and in different places. More than this, feminists showed that by slighting the discursive element of social welfare institutions we jeopardize our ability to understand the causal mechanisms that bring them into being. Without attending to the complex ways in which social policies are constructed as contested systems of meanings, scholars come to focus on measurable social processes or events (often exogenous to the welfare system itself) that are at best abstract and frequently inaccurate representations of what are essentially cultural systems — the underlying institutional logics that constitute the organizational field (Friedland and Alford, 1991).

The cultural turn is, however, far from complete. Feminists have enabled us to make a huge leap — but where we will land is still unclear. We need to get past the realist project for a science of welfare institutions, but does that imply that we should surrender the goal of explaining variations in welfare policies across time and place? We need to embrace a more interpretative stance and pursue thick descriptions of institutional processes but does that mean that we should abandon the tools of quantitative social science? My own suspicion is that the way forward, not just in the study of social welfare but in social science more generally, is to bring a more interpretative sensibility into the core of the social scientific project itself — to make science more interpretative and interpretation more scientific. To do this would be to fully accept the discursive orientation bequeathed to us by the feminist tradition and to do so in a way that also makes room for the very real advantages that can be gained from the incorporation of mathematics and technology —the two most important (and distinctive) elements of a formalist approach to scientific practice — into a new cultural sociology of institutional systems.

We have a long way to go before we can understand what such an endeavor would look like. Probably the change will require something like another paradigm shift, a movement away from the realist sensibilities of the post war era to a more genuinely constructivist social science (see Friedland and Mohr, 2004). My goal in this chapter is to take one concept, welfare discourse, and to suggest how we might advance a cultural sociology of the topic that embraces both interpretation and formal analysis. In particular I want to focus in the remainder of this chapter on two specific suggestions for how I think the formal study of discourse will affect the way we understand social institutions.

(1) Discourses have semiotic properties.

 

The semiotic component of institutional discourse that I want to focus on here concerns how meanings are constructed. A key analytic innovation that Saussure (1959) bequeathed to us was the concept of a synchronic analysis of meaning. Rather than understanding the meaning of a word by tracing its origins backwards in time and developing a narrative of becoming, Saussure sought to understand how word meanings were whole and complete in the moment. The key to synchronous interpretation was the recognition that word meanings (or sounds, or other semiotically defined system elements) are embedded within a discourse structure or a langue and thus meaning should be understood as a referential system constructed out of patterns of similarity and difference within a semiotic field. In Saussure’s formulation, “The content of a word is determined in the final analysis not by what it contains but by what exists outside it” (Saussure, 1959, p. 114). What followed from this insight was a long and hugely important tradition of interpretative work in anthropology and elsewhere that sought to explain the meaningfulness of things by understanding their position within a relational system of elements bound together in particular patterns of similarity and difference. This is the essence of structuralism.

Many scholars have sought to interpret the meaning of discourse systems in the field of social welfare. Indeed, most of the classic studies of welfare systems include some analysis of the classificatory distinctions that are embedded within the logic of the system, especially with regard to the ongoing problem of differentiating the worthy from the unworthy poor (e.g., Polanyi, 1957). These kinds of meaning systems can be fruitfully studied as semiotic fields and I think most scholars recognize this implicitly. My interest is in pushing harder on the interpretative technology of structuralism in order to further our ability to ground these interpretative acts in a systematic evaluation of discourse data. I can explain this point more clearly by discussing an example taken from my own research in which I make use of structuralist methods to analyze the discourse system of the 1907 New York City Charity Directory (Mohr, 1994).

First, I need to provide some context. The New York City Charity Directory was a large book published annually by the local Charity Organization Society, intended as a practical guide for relief workers in the city. It contained short (one or two paragraph) descriptions of nearly every organization operating in the field of social welfare during a given year. The directory was very comprehensive and it thus provides a window into what DiMaggio and Powell have described as an organizational field, “those organizations that, in the aggregate, constitute a recognized area of institutional life: key suppliers, resource and product consumers, regulatory agencies, and other organizations that produce similar services or products” (1983:148). I chose the 1907 directory because it was published at the height of the Progressive Era, a period during which fundamental changes were occurring in the field of social welfare.

Michael Katz (1986) describes this as a moment of transition between “the poorhouse era” and something that he labeled “the semi-welfare state.” The former designates the institutional system of relief that prevailed in American society during the nineteenth century. The poorhouse was the pivotal organization. This was a custodial facility, reminiscent of other custodial institutions that characterized the time — the insane asylum, the orphanage, the reformatory, and the penitentiary (Foucault 1979; Rothman 1971). Poorhouse inmates suffered an unpleasant fate and the unpleasantness was at least partially intended as a part of the logic of deterrence. The Progressive Era (including here the 10 or 15 years bracketing the turn of the century), marked the transition to a more modern institutional logic. The profession of social work was established, a modern set of scientific discourses were invented, the nonprofit sector proliferated, and organizations were rationalized in the Weberian sense of the term. It was during the Progressive Era that public discourse on poverty was softened a bit (especially when contrasted to the vilifications of the poor that characterized so much of the rhetoric of the 1880s) and intertwined with a proliferating set of discourses on social problems and “progressive” proposals for their solution.

Entries in the Charity Directory consisted of descriptions by each organization of their mission, their technologies and the types of individuals who were the objects of the organizations’ efforts. In this paper I focused on 15 of these classificatory designations (or status identities)—blind persons, consumptives, the disabled, ex-convicts, high-status individuals, immigrants, mothers, seamen, soldiers, strangers, tramps, the unemployed, unwed mothers, widows, and working people. Because of the theoretical importance of gender, I further divided each status identity into three subsets, those that were masculinized, those that were femininized and those that were left ungendered. This yielded a set of 38 distinct status identities.

In the paper I argue that we can learn about the meaning that these identity designations had within the discourse of social welfare if we formally treat them as elements within a semiotic system. To operationalize this concept I mapped the relations of similarity and difference among the identities by looking at how each class of individuals was treated in the welfare system (as represented discursively in this text). To facilitate this I divided organizational programs into 14 types of core activities that seemed to capture the range of relief technologies that were described in the Directory—giving money, giving food or other household necessities, providing work for pay, assistance in finding a job, temporary shelter, asylum (or long term shelter), incarceration (put in prison or in a reformatory), job training, domestic training (instruction on how to keep a proper home), counseling, religious direction, drug or alcohol (temperance) services, legal prosecution, vacation assistance, and community services. These were further subdivided according to the auspice of the organization performing the activity (public, religious, private nonprofit organization, etc.) yielding a set of 70 distinct treatment possibilities. I dropped all treatment combinations that were not applied to this set of identities and I also deleted all status identities that were not treated by one of these practices leaving a matrix of 26 identities and 58 practices.

The resulting rectangular matrix (which has identities in the columns and treatment practices in the rows) is composed of 0’s and 1’s (a 1 in a cell indicates that at least one organization in the 1907 Directory claimed to have treated the specified identity with a given practice). If we think of the identities as elements within a semiotic system we can use this matrix as a map of how each semiotic term (each status identity) is similar to or different from every other identity as defined by their treatment within the institutional field. Network analysts have provided a wide range of useful tools for analyzing problems of this sort. I make use of a technique here known as blockmodeling (White, Boorman and Breiger, 1976). This involves several steps. First the matrix of identities by practices is subjected to an ordinary product moment correlation analysis, the output matrix from the correlation is then itself subjected to a correlation analysis and this process is repeated iteratively until a clear bifurcation of the identities is achieved. Those new sub-groups are then each subjected to a separate analysis that produces another bifurcation and this is repeated as many times as desired (the procedure for doing this is known as CONCOR for convergence of iterated correlation analysis). The result is a hierarchical tree of bifurcated identity clusters (figure 1).

 

Figure 1. CONCOR Clustering of Status Identities into Structurally Equivalent Clusters,

1907 New York City Charity Directory

 

 

Identities within each of the 8 CONCOR clusters can be thought of as structurally equivalent to one another in the sense that they stand in comparable relations of similarity and difference to all the other identities in the field. Importantly, the clustering is not based on any essential attributes of people (their gender, their age, their wealth, etc.). Rather it is based on their relational identity, specifically their embeddedness within an institutional system of relief activities. Even here, however, the method embraces a relational over an essentialist logic in the sense that identities are not simply clustered together because they receive the same treatments or treatments that are substantively similar (e.g., public vs. private, punitive vs. non-punitive, physical vs. moral, etc.). Instead CONCOR puts identities in a common block location if they are subject to practices that are themselves relationally similar to one another, which is to say, practices that are applied to similar categories of people. In short, the method is very Sausserian; it defines entities purely according to their location within a system of relations.

Having parsed the identities into common structural locations with CONCOR, I proceed to blockmodel the clusters. Each cluster (block) is compared to every other by calculating the average of the correlations between all members of each set. This measure is constructed for every block pair and those pair-wise similarities that have values greater than average (.1375) are strongly related (as indicated by arrows in the blockmodel diagram, figure 2). Network analysts use this procedure to identify the relational locations or the “role structure” of individuals within a social network. I argue that the same approach when applied to an institutionally ordered textual field can be used to identify common discursive roles or what might also be thought of as shared structural locations within a moral order.

 

 

 

Figure 2. Blockmodel Structure of CONCOR Blocks,

1907 New York City Charity Directory

 

We know that social welfare systems are grounded in complex systems of exclusion and inclusion and that in Progressive Era New York these were based on a system of moral legitimations organized by gender. If our analytic strategy is effective, we should expect to see evidence of this in way in which both morality and gender are mapped onto the structural model. To test this assertion, I went back to the data and looked for all occurrences of status identities that were qualified by explicit moral descriptors. Thus whenever an identity in the Charity Directory is described by terms such as “deserving,” “innocent,” “respectable,” “worthy,” or alternatively as “degraded,” “depraved,” “dishonorable,” “profligate,” “sinful” (or the like) I coded that as an instance of an explicit moral evaluation. Figure 3 shows how the moral qualifications are distributed across the discourse role system. Strikingly, moral qualifications only occur in four of the 8 blocks, and the great majority of this discourse is restricted to blocks 3 and 5.

Figure 3. Interpretive Mapping of Block Structure,

1907 New York City Charity Directory

 

 

Looking at the blockmodel with these goals in mind (figure 3) suggests that there is utility in distinguishing between three major classes of identities. Tramps (a summary label for textual references to terms such as “rounders,” “vagrants,” “wanderers” and the like) and the gendered blind (block 8, which also contains references to those designated as “deaf,” “deaf-mute,” “dumb,” “defective” and so on) go together because they are structurally equivalent (both are structural isolates). I have labeled this group as being governed by a logic of “otherness.” A second group (blocks 2 and 3) is also connected by their sharing of structurally equivalent positions within the institutional discourse system and this group is definable by what would appear to be a logic of “exclusion”. It contains categories such as “strangers” (those excluded from the entitlement of community), the “disabled” (excluded from the entitlement of labor) and “unwed mothers” (excluded from the entitlement of domesticity). Block 3 includes “ex-convicts” (excluded from legality), the “unemployed” (excluded from the entitlement of labor but without the excuse of disability), and “female immigrants” (differentiated from other immigrants by the marking of their gender). Note that this block contains the highest proportion (31%) of discursive statements that contain explicitly articulated moral descriptors.

The third group includes all other blocks (4, 5, 6 and 7 — and note that 5 and 6 are collapsible as structurally equivalent). I have labeled this sector of moral space the “Logic of Entitlement” because it contains identities that were deemed to be legitimate recipients of social relief. Within this are three separate sub-logics of identity — membership within a guild community, within a class location or within a gendered sphere of domesticity. Many of the identities in block 4 (especially soldiers and seamen) are entitled to relief by virtue of their contributions as a member of an occupational group or as a war veteran. This is true of widows as well because their claims for relief are frequently linked to their husbands’ performance of duties in these domains. Immigrant identities represent a parallel type of status claim defined by co-membership within an ethnic community. In contrast, status identities in block 7 (mothers, working boys, girls, and women) would appear to be entitled to assistance as a result of their identification with the domestic sphere itself. Blocks 5 and 6 speak to an entitlement defined by identification with a specific class location, working men by virtue of their membership in the proletariat, high status individuals by virtue of their identification with the bourgeoisie. Notice that block 5 is the only other highly charged moral space within the field (24% of the references contain explicit moral qualification). This would seem to be a reflection of the inherent liminality of these more elite status identities within a moral order which performs the institutional function of differentiating among the categories of the poor.

Notice how gender and morality co-vary in this institutional space. Following Zerubavel’s (1993) concepts of marked and unmarked cultural categories, I have designated identities as gendered whenever their various forms are differentiated within the discourse role system (these are marked in bold face type). The blind, for example, are gendered (being spread across blocks 4 and 8), but HiStatus identities (all of which are in block 5) are not. By this reckoning, the moral space governed by the “Logic of Entitlement” is gendered (including 2 of 3 blocks and 10 out of 14 identities), the space of exclusion is ungendered and otherness is evenly split. I’ve put a vertical bar on the figure designating moral ambiguity. Those blocks of identities that are never linked to moral qualifications (1,2,6 and 8) are morally unambiguous. Those that require frequent moral specification (3 and 5) are highly ambiguous.

Together these insights suggest a hypothesis — moral entitlement covaries with gender, moral ambiguity does not. Given where we are in this endeavor, I would offer this up as an intriguing suggestion worthy of further investigation — no more than that. But I would also argue that the contribution here is not of any particular hypothesis, but the manner in which we have derived it. There is room for more interpretation and a clear need to explore these matters more extensively from a formal perspective. But leave that for the moment and let us turn to a second feature of welfare discourse that needs to be taken into consideration as we think about how to develop this style of cultural analysis.

 

(2) Discourses are mutually constitutive and dually ordered.

 

We can advance the development of a cultural science not only by plotting the semiotic features of these discourse systems but also by modeling welfare discourse as an element within a larger structural whole. Discourse systems are in no sense pure or self-contained. They are embedded in the social world and they have porous boundaries. Other discourse systems, other institutions, other elements of reality are articulated within the meanings of poverty relief, and indeed, are constitutive of them. In this section of the paper I want to focus on a second important feature of institutional discourse systems that I believe we must take into consideration — that they are mutually constitutive and dually ordered.

It is useful to see this problem in its intellectual context. One of the principal failings of traditional (French) structuralism was the problem of difference. What should count as difference? Most obviously in Derrida (1978) but also throughout the post-structuralist tradition scholars began to emphasize that no simple interpretative truths could be constructed but rather a multiplicity of interpretative possibilities. This rejection of epistemologies oriented toward objctive truth claims pushed the French intellectual field away from the more formalistic approaches to meaning towards post-modernism. Bourdieu was an exception. Though he well appreciated the limits of semiotics as an interpretative project he sought to hold onto a pragmatic scientism that worked with a post-structuralist sensibility rather than against it. Bourdieu argued that the social world was made up of a multiplicity of cultural discourse systems and that these were embedded in a given material reality manifested through a constellation of practices. The duality of this linkage, the connection between ways of knowing and ways of acting was accomplished in the phenomenological moment through something he called the habitus, and it was accomplished at the level of the institution in a social space he called a field.

The archetypical example of duality in Bourdieu (1977, 1990) is his discussion of the relationship that inheres between the material and the ideal world, the world of practice and the world of culture. Bourdieu’s famous essay on the Kabayle House (reprinted as an appendix to 1990) with its layered analysis of the simultaneous ordering of material space in terms of both practical household demands and a cosmological y rich tapestry of understandings is a case in point. Friedland and Alford (1991) suggest the example of the buying and selling of commodities as a set of practical activities that can only exist so long as people share a set of symbolic constructions that includes the idea of private property while at the same time the concept of property can only be truly meaningful in the context of a commodified world where market behavior is regularly conducted. An earlier generation of scholars called this a dialectic but it can also be understood as a particular type of structural form. Ron Breiger (1974) originally developed this concept (as well as his own set of mathematical tools for studying its manifestations) and defined it as a type of linkage in which two independent domains (such as culture and practice) are organized through systems of difference, yet neither order exists without the other because each constitutes the difference that exists within the other. Structural duality now serves as a pivotal construct in a promising new research program on the formal analysis of institutions (see for example, Breiger 2000, Mische and Pattison, 2000, Martin 2000, Harcourt, 2002).

I can best illustrate this concept by turning to another of my own papers, an article co-authored with Vincent Duquenne (Mohr and Duquenne, 1997). The paper also makes use of the Charity Directory dataset. This time, however, the data come from the 1888 Directory (the paper actually uses 4 years of data and perhaps its most important finding was that you can use these techniques to measure the degree of institutionalization over time, however, for the purposes of this discussion I will focus on just one year of data). The goal in this paper was to demonstrate a way to model the structural duality that theory told us we would find within the institutional logic of a field. Similar to before, the data are structured as a matrix of: (1) prevalent categories for naming/knowing the poor (these appear as column headers in table 1) and (2) institutionally recognizable repertoires of action for dealing with the poor (the row entries in table 1). The table is binary. A 1 in a cell entry indicates that (at least) one organization in the 1888 Charity Directory that employed a given (row) practice also employed a specific identity term (column) when referring to the object of that action. A 0 entry means that not a single organization in 1888 linked the practice with the identity.

 

Table 1. Poverty Practices by Poverty Categories (Binary)—1888

 

 

 

 

D

e

s

e

r

v

i

n

g

D

e

s

t

i

t

u

t

e

D

i

s

t

r

e

s

s

e

d

F

a

l

l

e

n

H

o

m

e

l

e

s

s

I

n

d

i

g

e

n

t

 

M

i

s

f

o

r

t

u

n

e

N

e

e

d

y

P

o

o

r

W

o

r

t

h

y

S

t

r

a

n

g

e

r

T

o

t

a

l

allocate money

give$

1

1

1

0

0

1

1

1

1

1

0

8

provide food

food

1

1

1

0

1

1

0

1

1

1

1

9

paid employment

paidWk

0

1

0

0

1

1

0

0

1

1

0

5

paid work in own home

homeWk

0

1

0

0

0

0

1

1

1

0

0

4

employment search

findJob

0

1

0

1

1

1

0

1

1

0

1

7

advise on work/family

advise

0

1

0

0

1

0

0

1

1

0

1

5

investigate home

investg

0

0

1

0

0

0

0

1

1

0

0

3

provide job training

jbTrain

0

1

0

0

0

0

0

1

1

0

0

3

give temporary shelter

shelter

1

1

0

1

1

1

0

1

1

1

1

9

long-term shelter

asylum

1

1

0

1

1

1

0

1

1

1

0

8

 

Total

4

9

3

3

6

6

2

9

10

5

4

61

 

The analysis is inspired by a simple assumption borrowed from Wittgenstein — the structure of meanings reflects the structure of use. A set theoretic analysis of table 1 appears in Figure 4. It shows the logical implications of the binary matrix as a system of sub-setting containments. A given term is below another if the set of practices applied to the former is a subset of the set of practices applied to the latter. The category “poor” has the broadest usage profile, suggesting that it conveys the most general concept of poverty. All other classifications can be seen as more specific refinements (sub-categories) of this general and all-embracing notion. Immediately below the poor is a bifurcation between the categories of the needy and the destitute, a split that would appear to capture a type of class-based differentiation, perhaps a reflection of the degree of severity of a person’s impoverishment. Other categories were either refinements of destitution (the indigent who were distinguished from the homeless) or a refinement of neediness (distressed presumably reflecting a more extreme or perhaps a more transient state of neediness). The only exception, the only class contradictory location among the terms indicating a category of the poor, was “misfortune” (a term that implied calamities of nature such as floods and earthquakes). Other terms were less clearly class specific. The moral designation of worthiness was restricted to being a refinement of the indigent, but the moral qualification of deservingness was a class contradictory extension of worthiness to the status of the needy. Strangers were also interstitial — partly needy, partly homeless.

A partial ordering of the practices (not reproduced here) can also be derived, but the real payoff comes from analyzing both of these orderings simultaneously in a single Galois lattice (figure 5). Because both orders are projected onto the same lattice structure (the smallest possible lattice in which the two can be embedded together), every point in the lattice represents the co-occurrence of the set of relief practices that are below it and the set of poverty categories that are above it. For clarity, the lattice is minimally labeled—a category is labeled at its highest occurrence, a practice is labeled at its lowest occurrence. Hence, the point labeled “Needy” is the highest point to which the category “Needy” applies. All points which fall on the lines descending from that point could also be labeled “Needy”. Reading from top to bottom, we can see the same sub-setting order that was observed in figure 4 (the destitute and the needy are subcategories of the poor, the distressed is a subcategory of needy, etc). The partial order of relief practices can also be retrieved from here. Reading from the bottom to the top (tracing the three lines ascending from the lowest point of the lattice), there are three first order relief practices—give$, give food and give shelter. The inverted triangle labeled “g:give$” marks the lowest point in the lattice in which the practice of giving money occurs. Subsets of practices then flow upward from here through the lattice. Thus, putting a person in an asylum is a practical subset of providing short-term shelter. The offer of paid work is a social practice that is institutionally located somewhere between (e.g., is a proper subset of) the practice of giving a person food and interring them in an asylum.

 

Figure 5. Lattice Analysis of Poverty Classifications and Practice Categories,

New York City Charity Directories, 1888

 

The great virtue of the Galois lattice is that it allows you to read these two orders against one another. While it is interesting to see that the practice of finding someone a job is (institutionally speaking) a subset of the practice of offering shelter, the really intriguing insight comes from seeing that the only difference between them is that the worthy and the deserving are both provided with shelter but not a single organization in the city of New York in 1888 publicly claimed to be trying to find work for worthy or deserving persons—try as you might, you will not be able to trace a line upwards in the lattice from Findjob to deserving or worthy. Apparently to be designated as worthy or deserving was to be located within the institutional order in such a way as to not be viewed as a candidate for regularized employment. By tracing out these kinds of cleavages (splits) in the lattice it becomes possible to discern some basic principles within the institutional system, principles that are dually ordered in the sense that they reflect co-variations (indeed, co-constitutions) in the classificatory logic of poverty specifications and in the practical repertoires of action within the field.

To cite just one more example discussed in the paper, consider again the basic bifurcation between the needy and the destitute. The lattice suggests that there are two relevant differences between these two social categories. One is whether a person was required to work in order to receive aid (paidWk). The other was whether a person would be the subject of a social investigation by a “friendly visitor. Clearly, these distinctions mattered. The demand for labor in exchange for relief was a more punitive approach to social welfare (in 1888 just as it is today) and was traditionally reserved for those classes of aid recipients who were regarded as less worthy. The lattice diagram suggests that this difference in treatment was a fundamental basis of differentiation between the destitute (who had to work) and the needy (who did not).

While it was largely punitive, the requirement that relief applicants perform some labor in exchange for assistance was also a mechanism for knowing the poor. The symmetry here is quite informative. Social investigations were, in a sense, a more modern (and more rationalized) mechanism for accomplishing the same thing. Friendly visitors (proto-social workers) were dispatched to assess an applicants' moral character, social habits, housekeeping skills and parenting practices in order to help diagnose the social disease (Richmond, 1917). Thus, one could say that, in 1888, the needy and the destitute were primarily distinguished by the modality of surveillance or (in Foucault’s terms) the regime of power/knowledge that they were subjected to and which mediated their relationship to the institutional field of social welfare. Those classes of the poor that were expected to demonstrate their moral fortitude and economic desperation by passing a “labor test” were classified as destitute. Those classes of the poor that were subjected to the more modern, less physical (though hardly less demeaning) requirement that they subject themselves to a social investigation were classified as needy.

In the paper we provide a full reading of the lattice (and its variations across time) and contend that such a reading allows us to provide an exegesis of this cultural system that resonates with extant theoretical expectations regarding the dialectical (or dualistic) ordering of culture and practice. But more than this, we argue that the analysis allows us to bring the tools of formal analysis to bear on what are essentially interpretive problems, enabling us to construct ways of seeing the residue of historical events that highlight and bring into relief semiotic features of the institutional order that resonate with and complement other types of interpretative understandings. In short, these are methods that provide a style of knowing that can supplement and interact with more traditional readings, a style of research that has the added advantage of being replicable, consistent with hypothesis testing methodologies and well suited to styles of analytic specification that facilitate the identification of structural reductions and deeply ordered logical sub-systems within the discourse itself.

 

Conclusion

 

            The shift from a realist to a constructionist perspective in the study of social welfare is well under way. Feminist scholars inaugurated the change by calling into question taken for granted categories of analysis and by insisting on the social productivity of discourse. But the sociological study of discourse is still young, and we have far to go. I have suggested two analytic principles that strike me as being particularly worthy of further pursuit: (1) discourse is semiotically ordered and amenable to relational analysis and (2) discourse systems are interpenetrated (and constituted) by their articulations within systems of practice and this makes them amenable to analytic strategies that reveal dualities within multi-ordered relational systems.

By now it is surely apparent that another of my concerns is to develop a more formally grounded approach to understanding systems of cultural discourse. In closing I want to quickly comment on what I think the role of science can and should be in such an endeavor. Science is really about two kinds of things. On the one hand, it is about a particular type of professionalization and rationalization of the knowledge production process — about the ways in which communities of scholars are organized, about how they are taught to ask questions and how the scientific method is to be applied. A lot of ink has been spilled over these qualities of scientific life, not all of it productively. But science is also about the use of technology in the production of knowledge and this is the part most relevant to the current discussion.

Astronomy strikes me as a useful example for thinking about the role of technology in science. Astronomers conduct their investigations on the basis of information compiled from sophisticated signal detection equipment created to measure types of wave particles that the human senses are incapable of perceiving. Beautiful images of the distant universe, iconic representations of the cosmos (think of recent news magazine covers) are essentially aesthetically rendered statistical summaries of these data-streams. Astronomers use these images (and analyses of the statistical systems that underlie them) as a way to construct an informed community dialogue about what might be going on out there. Astronomers’ relationships to this dialogue, and the instruments that enable it, are wholly human (as science studies scholars have assured us) which means that we should not be too persuaded by those who would claim that true science is defined by its privileged relationship to objective reality. Astronomers and other natural scientists are no less tied to the plodding and all too human project that characterizes institutional life. But what astronomers do have—as I also have, sitting here at my computer—is a technical system that is productive in the sense that it substantively facilitates the endeavors we pursue.

My suggestion is that cultural sociology should also invite technology in. Though they are riddled with limitations, scientific instruments are nonetheless useful in furthering the pursuit of human ends because they can extend the limitations of our physical selves. Like the astronomers, we should put our machines to work sifting through streams of data taken from the textual universe. And, like the astronomers, we should work with these machines, tinker with them, coax them to become ever more effective signal rendering devices. My sense is that we can gain a great deal from this endeavor because there is so much more textual information out there than we can possibly hope to perceive as an embodied human reader. I have tried through these examples to suggest some of the features of institutional discourse that I think a more scientific (e.g., a signal collection, analysis and enhancement) approach to interpretation ought to consider.

            Limits of time and space prevent me from taking up a variety of other equally important issues. In particular, my rather single minded focus here on laying out some strategies for discourse analysis has left aside all of the broader questions of how it is that discursive processes are themselves dually articulated at the institutional level with broader systems of social organization (including the logic and character of organizational fields) and at the subjective level with phenomenologically experienced systems of habitus, subjectivity and tactical agency. These are matters that must be tackled head on if we are to make good on the promise of developing a fuller science of cultural discourse.

 

Bourdieu, Pierre. 1977. Outline of a Theory of Practice. Cambridge: Cambridge University Press.

Bourdieu, Pierre. 1990. The Logic of Practice. Stanford: Stanford University Press.

Breiger, Ronald L. 1974. "The Duality of Persons and Groups." Social Forces 53: 181-90.

Breiger, Ronald L. 2000. “A Tool Kit for Practice Theory.” Poetics. Vol. 27/2-3:91-116.

Derrida, Jacques 1978 [1967]Writing and Difference. University of Chicago Press: Chicago.

DiMaggio, Paul J. and Walter W. Powell. 1983. "The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields." American Sociological Review 48: 147-160.

Esping-Anderson, GŅsta and Walter Korpi. 1984. "Social Policy as Class Politics in Post-War Capitalism: Scandinavia, Austria, and Germany." Pp. 179-208 in Order and Conflict in Contemporary Capitalism, edited by J.H. Goldthorpe. Oxford: Oxford University Press.

Esping-Anderson, GŅsta. 1990. The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press.

Foucault, Michel. (1975) 1979. Discipline and Punish: The Birth of the Prison. New York: Vintage Books.

Fraser, Nancy. 1989. Unruly Practices: Power, Discourse and Gender in Contemporary Social Theory. Minneapolis: University of Minnesota Press.

Friedland, Roger and John W. Mohr. Forthcoming. 2004. “The Cultural Turn in American Sociology.” Matters of Culture: Cultural Sociology in Practice, Roger Friedland and John Mohr (eds). Cambridge University Press.

Friedland, Roger and Robert R. Alford. 1991. "Bringing Society Back In: Symbols, Practices and Institutional Contradictions." Pps. 232-263 in The New Institutionalism in Organizational Analysis edited by Walter W. Powell and Paul DiMaggio. Chicago: University of Chicago Press.

Furner, Mary O. 1975. Advocacy and Objectivity: A Crisis in the Professionalization of American Social Science, 1865-1905. Lexington, Kentuky: The University Press of Kentucky.

Geertz, Clifford. 1973. "Thick Description: Toward an Interpretive Theory of Culture." Pp. 3-30 in The Interpretation of Cultures. New York: Basic Books.

Gordon, Linda 1990. "The New Feminist Scholarship on the Welfare State." Pp. 9-35 in Women, the State and Welfare edited by Linda Gordon. Madison: University of Wisconsin Press.

Harcourt, Bernard. 2002. “Measured Interpretation: Introducing the Method of Correspondence Analysis to Legal Studies.” University of Illinois Law Review: 2002(4):979-1018.

Hasenfeld, Yeheskel, Jane A. Rafferty and Mayer N. Zald. 1987. "The Welfare State, Citizenship and Bureaucratic Encounters." Annual Review of Sociology 13: 387-415.

Katz, Michael B. 1986. In the Shadow of the Poorhouse: A Social History of Welfare in America. New York: Basic Books.

Martin, John Levi. 2000a. “What Do Animals Do All Day?: The Division of Labor, Class Bodies, and Totemic Thinking in the Popular Imagination.” Poetics. Vol. 27/2-3:195-231.

Merton, Robert K. 1957. "The Sociology of Knowledge." Pps. 456-488 in Social Theory and Social Structure. Glencoe: Free Press.

Mische, Ann and Philippa Pattison, 2000. “Composing a Civic Arena: Publics, Projects, and Social Settings.” Poetics. Vol. 27/2-3:163-194.

Mohr, John W. 1994. "Soldiers, Mothers, Tramps and Others: Discourse Roles in the 1907 New York City Charity Directory." Poetics 22, pp. 327-357.

Mohr, John W. 1998. "The Classificatory Logics of State Welfare Systems: Towards a Formal Analysis." Pages 207-38 in Public Rights, Public Rules: Constituting Citizens in the World Polity and National Policy, Connie McNeely (ed.). NY: Garland Publishing, Inc.

Mohr, John W. Forthcoming. "Implicit Terrains: Meaning, Measurement, and Spatial Metaphors in Organizational Theory." Forthcoming in Constructing Industries and Markets, Jospeh Porac and Marc Ventresca (eds). NY: Elsevier

Mohr, John W. and Vincent Duquenne. 1997. “The Duality of Culture and Practice: Poverty Relief in New York City, 1888-1917.” Theory and Society, (April/June) Vol. 26/2-3: 305-356.

O'Connor, James. 1973. The Fiscal Crisis of the State. New York: St. Martin's Press.

Polanyi, Karl. [1944] 1957. The Great Transformation: The Political and Economic Origins of Our Time. Boston: Beacon Press.

Ross, Dorothy. 1991. The Origins of American Social Science. New York: Cambridge University Press.

Rothman, David J. 1971. The Discovery of the Asylum: Social Order and Disorder in the New Republic. Boston: Little, Brown and Company.

Saussure, Ferdinand de. 1959. Course in General Linguistics. New York: McGraw-Hill

Skocpol, Theda and John Ikenberry. 1983. "The Political Formation of the American Welfare State in Historical and Comparative Perspective." Comparative Social Research 6: 87-148.

Thomas, George M., John W. Meyer, Francisco O. Ramirez, and John Boli. 1987. Institutional Structure: Constituting State, Society and the Individual. Beverly Hills: Sage.

White, Harrison C., Scott A. Boorman and Ronald L. Breiger. 1976. "Social Structure From Multiple Networks. I. Blockmodels of Roles and Positions." American Journal of Sociology 81: 730-780.

Wilensky, Harold L. and Charles N. Lebeaux. (1958) 1965. Industrial Society and Social Welfare: The Impact of Industrialization on the Supply and Organization of Social Welfare Services in the United States. New York: Free Press.

Zerubavel, Eviatar. 1993. The Fine Line: Making Distinctions in Everyday Life. Chicago, IL: University of Chicago Press.