Saturday, February 27, 2016

Values, directions, and action


Several earlier posts have raised the question of rational life planning. What is involved in orchestrating one's goals and activities in such a way as to rationally create a good life in the fullness of time?

We have seen that there is something wildly unlikely about the idea of a developed, calculated life plan. Here is a different way of thinking about this question, framed about directionality and values rather than goals and outcomes. We might think of life planning in these terms:
  • The actor frames a high-level life conception -- how he/she wants to live, what to achieve, what activities are most valued, what kind of person he/she wants to be. It is a work in progress.
  • The actor confronts the normal developmental issues of life through limited moments in time: choice of education, choice of spouse, choice of career, strategies within the career space, involvement with family, level of involvement in civic and religious institutions, time and activities spent with friends, ... These are week-to-week and year-to-year choices, some more deliberate than others.
  • The actor makes choices in the moment in a way that combines short-term and long-term considerations, reflecting the high-level conception but not dictated by it.
  • The actor reviews, assesses, and updates the life conception. Some goals are reformulated; some are adjusted in terms of priority; others are abandoned.
This picture looks quite a bit different from more architectural schemes for creating and implementing a life plan considered in earlier posts, including the view that Rawls offers for conceiving of a rational plan of life. Instead of modeling life planning after a vacation trip assisted by an AAA TripTik (turn-by-turn instructions for how to reach your goal), this scheme looks more like the preparation and planning that might have guided a great voyage of exploration in the sixteenth century. There were no maps, the destination was unknown, the hazards along the way could only be imagined. But there were a few guiding principles of navigation -- "Keep making your way west," "Sail around the biggest storms," "Strive to keep reserves for unanticipated disasters," "Maintain humane relations with the crew." And, with a modicum of good fortune, these maxims might be enough to lead to discovery.

This scheme is organized around directionality and regular course correction, rather than a blueprint for arriving at a specific destination. And it appears to be all around a more genuine understanding of what is involved in making reflective life choices. Fundamentally this conception involves having in the present a vision of the dimensions of an extended life that is specifically one's own -- a philosophy, a scheme of values, a direction-setting self understanding, and the basics needed for making near-term decisions chosen for their compatibility with the guiding life philosophy. And it incorporates the idea of continual correction and emendation of the plan, as life experience brings new values and directions into prominence.

The advantage of this conception of rational life planning is that it is not heroic in its assumptions about the scope of planning and anticipation. It is a scheme that makes sense of the situation of the person in the limited circumstances of a particular point in time. It doesn't require that the individual have a comprehensive grasp of the whole -- the many contingencies that will arise, the balancing of goods that need to be adjusted in thought over the whole of the journey, the tradeoffs that are demanded across multiple activities and outcomes, and the specifics of the destination. And yet it permits the person to travel through life by making choices that conform in important ways to the high-level conception that guides him or her. And somehow, it brings to mind the philosophy of life offered by those great philosophers of life, Montaigne and Lucretius.

Thursday, February 25, 2016

Guest post by Gianluca Pozzoni on political entities

Gianluca Pozzoni is a PhD Candidate in Political Studies at the University of Milan, Italy. His interests span the foundations of the social sciences, and he has written on Marxism, methodological individualism, and the status of social structures. Thank you, Gianluca, for contributing this stimulating guest post.


Daniel Little’s recent post on Assemblage theory as heuristic raises important and thought-provoking issues for social theorizing.

As Little understands it, Manuel DeLanda’s theory put forth in A New Philosophy of Society can be seen as a manifesto for non-essentialist and non-reductionist social theories. In this view, social entities do not have a fixed place in a vertical hierarchy that moves from the building blocks of the social world (e.g. agents) all the way up to large-scale structures (e.g. global markets).

Instead, the components of the social world differ according to the “assemblages” under consideration. So, for instance, organizations are assemblages of people; nation-states are assemblages of cities, people, and organizations; cities are assemblages of people, organizations, as well as buildings and infrastructures; and so on. Thus understood, assemblage theory allows for no basic constituents of the social because it does away with the very idea of society as a structured totality with a “basis” and a “summit”.

This seems to go in the direction of a “flat social ontology” as described by Daniel Little here. A flat model of the social reality would position actors and their interaction networks (e.g. organizations) all at the same level, without further assumptions as to how some causal powers are “emergent” on others (if at all).

This perspective has undeniable attractiveness. For one thing, it is theoretically parsimonious. It does away with somewhat obscure or ill-defined notions of “emergence” while at the same time allowing for aggregate social entities to be irreducible to their components, for instance by assuming that they have at least an independent functioning. Furthermore, it does not presuppose an abstract layering of the social world, let alone one metaphysically pre-ordered into “higher” and “lower” levels.

Nonetheless, Little has already detailed some limitations of the flat-ontology perspective. For Little, conceiving social reality as flat ultimately boils down to assuming an ontology like the one associated with spare versions of methodological individualism.

Here I wish to point out a different problematic aspect of this approach.

Even setting aside the depth ordering of social aggregates, social theory often makes use of broad categories that help group together heterogeneous classes of phenomena for explanatory purposes. These are, for instance, “the polity”, “the economy”, “the military”, and the like.

What ontological status do these categories have? Let us consider politics. In his Contentious Performances (2008), Charles Tilly writes:
We enter the realm of politics when we interact with agents of governments, either dealing with them directly or engaging in activities bearing on governmental rights, regulations, and interests. (p. 6)
However restrictive this definition may be, it is able to identify the political domain in terms of the typology of networks involved – namely, governmental institutions. Accordingly, Tilly et al. (2001) could use it to group together and explain phenomena as diverse as the Watergate scandal and the Mau Mau rebellion in 1950s Kenya under the common label of “contentious politics”.

In a layered social ontology, governmental institutions can be seen as making up a level of the social in its own right. As Little puts it:
For example, might the state be a level-2 entity, in that it encompasses organizations and individuals and it possesses new causal properties not present at level 1? In principle this seems possible. The state is a complex network of organizations and individuals. And it is logically possible that new causal powers emerge that depend on both base and level 1, but that do not require reduction to those lower-level properties.
If this can be granted, politics can be ontologically grounded onto a level of the social. Specifically, we may confine political ontology to all phenomena that involve government-level causal powers.

Flat social ontologies, however, reject the very notion of “levels” of the social. At best, compound social entities exist as relations among their components; they may well have an independent functioning (link), but no “higher-level” or “emergent” causal powers. So do states, governments, and the like.

What then would a flat political ontology look like? Perhaps it may be helpful to refer to Adrian Leftwich’s distinction in What is Politics: The Activity and its Study between politics as “process” and politics as “arena”:
The latter, or arena, approach tends to have a narrower and sharper focus (normally the state and the institutions of government and local government – sometimes, in a more comparative context, including kings, chiefs or emperors and their courts and their relations with the public). (p. 13)
The theory of politics as arena seems to fit quite nicely the stratified approach to social ontology. The arena metaphor suggests that governmental institutions – national or otherwise – make up a relatively autonomous set of interactions quite distinct from other forms of interaction. This can be supported from an ontological point of view by assuming that such institutions are endowed with causal powers that are all situated at one specific level.

Leftwich contrasts the arena model of politics with the idea of politics as “a general process [or set of processes – G.P.] which is not confined to certain institutional arenas or sites” (p. 14). This dynamic idea may seem appealing to those who do not wish to build political ontology as a “regional ontology”. The processual definition of politics does not make reference to the causal powers of any specific entities and therefore does not necessarily require an ontological layering of them.

However, the obvious question then arises as to what makes processes political as opposed to, say, economic. This question seems to be of crucial relevance if the category of politics is to serve explanatory purposes. As Leftwich puts it:
… does such an encompassing view mean that every human interaction is political in some respect? If so, and if politics is thus so broadly defined, what is left that is distinctive about it? (p. 14)
A catch-all definition of politics, then, seems to be redundant and to have little explanatory use. Unfortunately, the flat theory of society alone hardly provides any guidance for further theorizing along this line.

[Query from Dan Little in response to the final question from Leftwich: Is it possible that the difference between economic and political processes is not after all an ontological difference, but rather a pragmatic difference of classification for us? In other words, it is not a fact about the world but about our interests that "constitution" and "wage labor" are classified as "political/legal" and "economic".] 

Friday, February 19, 2016

Causal diagrams and causal mechanisms

There is a long history of the use of directed causal diagrams to represent hypotheses about causation. Can the mathematics and graphical systems created for statistical causal modeling be adapted to represent and evaluate hypotheses about causal mechanisms and outcomes?

In the causal modeling literature the structure of a causal hypothesis is something like this: variable T increases/ decreases the probability of the occurrence of outcome E. This is the causal relevance criterion described by Wesley Salmon in Scientific Explanation and the Causal Structure of the World. It is a fundamentally statistical understanding of causality.

Here is a classic causal path model by Blau and Duncan indicating the relationships among a number of causal factors in bringing about an outcome of interest -- "respondent's first job".

This construction aims at joining a qualitative hypothesis about the causal relations among a set of factors with a quantitative measurements of the correlations and conditional probabilities that support these causal relations. The whole construction often takes its origin in a multivariate regression model.

Aage Sørensen describes the underlying methodological premise of quantitative causal research in these terms in his contribution to Frontiers of Sociology (Annals of the International Institute of Sociology Vol. 11):
Understanding the association between observed variables is what most of us believe research is about. However, we rarely worry about the functional form of the relationship. The main reason is that we rarely worry about how we get from our ideas about how change is brought about, or the mechanisms of social processes, to empirical observation. In other words, sociologists rarely model mechanisms explicitly. In the few cases where they do model mechanisms, they are labeled mathematical sociologists, not a very large or important specialty in sociology. (370)
My question here is whether this scheme of representation of causal relationships and the graphical schemes that have developed around it are useful for the analytics of causal mechanisms.

The background metaphysics assumed in the causal modeling literature is Humean and "causal-factor" based; such-and-so factor increases the probability of occurrence of an outcome or an intermediate variable, the simultaneous occurrence of A and B increase the probability of the outcome, etc. Quoting Peter Hedstrom on causal modeling:
In the words of Lazarsfeld (1955: 124-5), "If we have a relationship between x and y; and if for any antecedent test factor the partial relationships between x and y do not disappear, then the original relationship should be called a causal one." (Dissecting the Social: On the Principles of Analytical Sociology)
The current iteration of causal modeling is a directed acyclic graph (DAG). Felix Elwert provides an accessible introduction to directed acyclic graphs in his contribution to Handbook of Causal Analysis for Social Research (link). Here is a short description provided by Elwert:
DAGs are visual representations of qualitative causal assumptions: They encode researchers’ expert knowledge and beliefs about how the world works. Simple rules then map these causal assumptions onto statements about probability distributions: They reveal the structure of associations and independencies that could be observed if the data were generated according to the causal assumptions encoded in the DAG. This translation between causal assumptions and observable associations underlies the two primary uses for DAGs. First, DAGs can be used to prove or disprove the identification of causal effects, that is, the possibility of computing causal effects from observable data. Since identification is always conditional on the validity of the assumed causal model, it is fortunate that the second main use of DAGs is to present those assumptions explicitly and reveal their testable implications, if any. (246)
A DAG can be interpreted as a non-parametric structural equation model, according to Elwert. (Non-parametric here means simply that we do not assume that the data are distributed normally.) Elwert credits the development of the logic of DAGs to Judea Pearl and Peter Spirtes, along with other researchers within the causal modeling community.

Johannes Textor and a team of researchers have implemented DAGitty, a platform for creating and using DAGs in appropriate fields, including especially epidemiology (link). A crucial feature of DAGitty is that it is not solely a graphical program for drawing graphs of possible causal relationships; rather, it embodies an underlying logic which generates expected statistical relationships among variables given the stipulated relationships on the graph. Here is a screenshot from the platform:

The question to consider here is whether there is a relationship between the methodology of causal mechanisms and the causal theory reflected in these causal diagrams. 

It is apparent that the underlying ontological assumptions associated with the two approaches are quite different. Causal mechanisms theory is generally associated with a realist approach to the social world, and generally rejects the Humean theory of causation. The causal diagram approach, by contrast, is premised on the Humean and statistical approach to causation.  A causal mechanisms hypothesis is not fundamentally evaluated in terms of the statistical relationships among a set of variables; whereas a standard causal model is wholly intertwined with the mathematics of conditional correlation.

Consider a few examples. Here is a complex graphical representation of a process understood in terms of causal mechanisms from McGinnes and Elandy, "Unintended Behavioural Consequences of Publishing Performance Data: Is More Always Better?" (link):

Plainly this model is impossible to evaluate statistically by attempting to measure each of the variables; instead, the researchers proceed by validating the individual mechanisms identified here as well as the direction of influence they have on other intermediate outcomes. The outcome of interest is "quality of learning" at the center of the graph; and the diagram attempts to represent the complex structure of causal influences that exist among several dozen mechanisms or causal factors.

Here is another example of a causal mechanisms path diagram, this time representing the causal system involved in drought and mental health by Vins, Bell, Saha, and Hess (link).

Here too the model is not offered as a statistical representation of covariance among variables; rather, it is a hypothetical sketch of the factors which play in mechanisms leading from drought to depression and anxiety in a population. And the assessment of the model should not take the form of a statistical evaluation (a non-parametric structural equation model), but rather a piecemeal verification of the validity of the specific mechanisms cited. (John Gerring argues that this is a major weakness in causal mechanisms theory, however, in "Causal Mechanisms? Yes, But ..." (link).)

It seems, therefore, that the superficial similarity between a causal model graph (a DAG) and a causal mechanisms diagram is only skin-deep. Fundamentally the two approaches make very different assumptions about both ontology (what a causal relationship is) and epistemology (how we should empirically evaluate a causal claim). So it seems unlikely that it will be fruitful for causal-mechanisms theorists to attempt to adapt methods like DAGs to represent the causal claims they want to advance and evaluate.

Thursday, February 11, 2016

Assemblage theory as heuristic

In A New Philosophy of Society: Assemblage Theory and Social Complexity Manuel DeLanda takes up one of Deleuze's key ideas. This is the idea of "assemblage", and it has been discussed here several times previously (link). (See DeLanda's extensive EGS lecture on assemblage theory below.) Here is a preliminary discussion of assemblage in New Philosophy of Society.
Today, the main theoretical alternative to organic [Hegelian] totalities is what the philosopher Gilles Deleuze calls assemblages, wholes characterized by relations of exteriority. These relations imply, first of all, that a component part of an assemblage may be detached from it and plugged into a different assemblage in which its interactions are different. In other words, the exteriority of relations implies a certain autonomy for the terms they relate, or as Deleuze puts it, it implies that 'a relation may change without the terms changing'. Relations of exteriority also imply that the properties of the component parts can never explain the relations which constitute a whole, that is, 'relations do not have as their causes established ...' although they may be caused by the exercise of a component's capacities. In fact, the reason why the properties of a whole cannot be reduced to those of its parts is that they are the result not of any aggregation of the components' own properties but of the actual exercise of their capacities. These capacities do depend on a component's properties but cannot be reduced to them since they involve reference to the properties of other interacting entities. Relations of exteriority guarantee that assemblages may be taken apart while at the same time allowing that the interactions between parts may result in a true synthesis. (10-11)
In addition to the exteriority of relations, the concept of assemblage is defined along two dimensions. One dimension or axis defines the variable roles which an assemblage's components may play, from a purely material role at one extreme of the axis, to a purely expressive role at the other extreme.... The other dimension defines variable processes in which these components become involved and that either stabilize the identity of an assemblage, by increasing its degree of internal homogeneity or the degree of sharpness of its boundaries, or destabilize it. (12)
In an illuminating discussion of some of Fernand Braudel's comments about medieval villages, DeLanda writes:
This brief description yields a very clear picture of a series of differently scaled assemblages, some of which are component parts of others which, in turn, become parts of even larger ones. (18)
What does this mean in practical terms? As a first approximation, the core idea of assemblage is that social things (cities, structures, ideologies) are composed of an overlapping and contingent collection of a heterogeneous set of social activities and practices. The relations among these activities and practices are contingent, and the properties of the composite thing -- the assemblage -- are likewise a contingent and "emergent" sum of the properties of the component threads. The composite has no "essence" -- just a contingent and changeable set of properties. Here is the thumbnail description I provided in the earlier post:
Fundamentally the idea is that there does not exist a fixed and stable ontology for the social world that proceeds from "atoms" to "molecules" to "materials". Rather, social formations are assemblages of other complex configurations, and they in turn play roles in other, more extended configurations. (link)
Here I want to ask a very simple preliminary question: What is the intellectual role of assemblage theory for sociology and for the philosophy of social science? Is assemblage theory a substantive social theory? Is it a guide to research and methodology? Or is it an ontology?

I think we do best to understand assemblage theory as a high-level and abstract ontological framework, an abstract description of the nature of the social world. It highlights the pervasive fact of  the heterogeneous nature of phenomena in the social world. But it does not provide a substantive theory of what those component threads are; this is for concrete sociological theory to work out. Unlike rational choice theory, Marxist theory, or pragmatist action theory -- each of which rests upon a substantive core set of ideas about the fundamentals of social action and structure -- assemblage theory is neutral with respect to these topics.

So assemblage theory is not a guide to the constituents of the social world; it is not similar to atomic theory or the Mendeleev table of the elements. However, I believe the theory is indeed methodologically helpful. Exploring assemblage theory is a potentially valuable activity for social scientists and philosophers. This is because the theory encourages us to study component systems and underlying social processes rather than looking for unified theories of large unified social objects. In this way it gives value and direction to multi-theoretical, inter-disciplinary approaches.

Moreover, this approach encourages social scientists to arrive at partial explanations of social features by discerning the dynamics of some of the components. These accounts are necessarily incomplete, because they ignore many other constituents of the assembled whole. And yet they are potentially explanatory, when the dynamics being studied have the ability to generate trans-assemblage characteristics (continuity, crisis). (This seems to have some resonance with Roy Bhaskar's idea that the social world is an "open" system of causation; A Realist Theory of Science.)

So assemblage theory is not a substantive social theory. It doesn't prescribe any specific ideas about the components, layers, laminations, or threads out of which social phenomena are composed. Instead, it offers a vision of how we should think of all such constructions in the social world. We should be skeptical about the appearance of unity and coherence in an extended social entity (e.g. the Justice Department or the Muslim world), and look instead to discover some of the heterogeneous and independent processes that underlie the surface appearance. And it gives ontological support for some of the theoretical inclinations of comparative historical sociology (Tilly, Steinmetz, Mann): look for the diversity of social arrangements and the context-dependent conjunctural causes that underlie complex historical events.

Here is a lecture by DeLanda on assemblage theory.

(I chose the illustration of a circus at the top because a circus illustrates some of the layered compositionality that assemblage theory postulates: multiple agents playing multiple roles; transportation activities and business procedures; marketing ploys and aesthetic creativity; and many things happening in the three rings at the same time.)

Sunday, February 7, 2016

Meanings and mechanisms

image: photographs of Martin Luther King, Jr. at the University of Michigan, 1962

There are two large categories of factors that are fundamental to understanding social processes -- meanings and mechanisms. I’ve given a preponderance of attention to the importance of social causal mechanisms within historical and social explanation (link). We explain a social outcome when we identify the social mechanisms that brought it about.

It is crucial to bear in mind always, however, that there is a complementary dimension to social life and social process — the pervasive fact that people act within frames of meanings and interpretations that they bring to their social relationships and their social worlds. Human action is meaningful action, and we can't make sense of action without attributing meanings, intentions, and frameworks of understanding and desire to the individuals who constitute a social encounter. 

This is not a new insight, of course; it was fundamental to the hermeneutic approach to social life, including the influential thinking of Wilhelm Dilthey (Introduction to the Human Sciences). But the classical hermeneutic approach tended to under-value the importance of causation and mechanisms in the social world; whereas it is clear today that both mechanisms and meanings are inseparably embedded within the social world.

It is in fact misleading to portray mechanisms and meanings as complementary “dimensions” of social change. Rather, we might say that mechanisms depend upon meanings, for the simple reason that mechanisms depend upon actions, and actions presuppose meanings. This is the thrust of my emphasis on "actor-centered" approaches to sociology (link). The actor-centered perspective takes seriously the meanings, values, cognitive and practical frameworks that individuals bring to their interactions in the social world, and it urges social scientists and historians to improve upon their current theories of the actor.

Institutions and organizations are often invoked as causal factors or mechanisms in the production of important social outcomes. But institutions always work by influencing the behavior of the individual actors whom they touch; so either explicitly or implicitly we need to have a theory of the actor's mental frameworks if we are to understand the causal power of institutions to influence outcomes. 

If we want to know why there is grade inflation in universities, we need to refer to some of the institutional mechanisms that influence grading practices (causal influences!), but we also need to refer to the goals and meanings that participants bring to the interaction between students, professors, and appeals committees. Sometimes those mental frameworks are trivial and manifest -- students want higher grades for reasons of career success as well as personal validation, faculty want to function in accordance with their responsibilities as neutral assessors of academic performance while at the same time demonstrating empathy for the needs of their students. These interlocking intentions and desires lead to a dynamic movement of average grades over time -- sometimes higher, once in a while lower. But sometimes the underlying mental frameworks that drive important social outcomes are more obscure -- for example, the disaffection and doubt that leads inner city minority students to despise high school. 

Think for a moment about how meanings and intentional actions give rise to a common social mechanism, hate-based nationalist mobilization. A few strident leaders formulate a message of hate against a group -- currently, MENA immigrants in various European countries; they find means of gaining access to national media (through provocative demonstrations); and they extend their influence from the tiny percentage of racist extremists ex ante to a sizable percentage of the more moderate population. How does this work? Why do ordinary non-racist citizens fall prey to the hateful messages of the extreme right? Presumably a convincing answer will depend on the specifics of the communications strategies and messages conveyed by the nationalist party, interlocking with an astute reading of the fears and suppressed prejudices of the majority population. In other words, the mechanism of racist mobilization depends on a substratum of political emotion and belief that can be adroitly manipulated by the racist group and its leaders.

Philosophers sometimes distinguish meanings and causes as subjective and objective.(This is implied in Georg Henrik von Wright's classic book Explanation and Understanding.) But this is not a useful way of thinking about the two categories. Meanings are often fully objective -- in the sense that we can investigate them empirically and they can be demonstrated to have stable and enduring effects in the world. And social causes have an element of subjectivity built into them, for the simple reason that social causes always invoke the subjective states of mind of the actors who make them up. It isn't even accurate to say that meanings exist solely within the actor, whereas causes exist outside the actor. The meanings that Weber identified in the notion of the Protestant Ethic are indeed embodied in a population of individuals (inward); but they are pervasive and influential on those same individuals (outward). So the Protestant Ethic is both an inner state of mind and an external and coercive set of values and beliefs.

(The photos of Dr. King above are relevant in this context because the civil rights struggle of the 1950s and 1960s offers ample examples of meanings and mechanisms in the evolving mobilizations, legislation, cycles of Jim Crow violence, and emerging ideas about Black Power within the African-American community.)

Wednesday, February 3, 2016

China's developmental resettlements

The human costs of China's Three Gorges dam project are reasonably well known. Since construction began in 1994 between 1.3 million and two million people have been involuntarily resettled to higher ground and to other provinces. The project has created massive environmental hazards for China (link), and has also created a gigantic human cost among the families and communities who were forcibly relocated.

Less well known are earlier waves of "hydraulic refugees" along the Yellow River beginning with construction of the Sanmenxia Dam in 1954 (completed in 1960). And subsequent water control projects through the next several decades have created yet other large-scale displacements along the Yellow River across north-central China. The movements of peoples created by these major projects of post-Revolution-era civil and social engineering have continued to reverberate until the present.

image: Yellow River topography

In the Sanmenxia Dam case the total number of forced migrants is probably lower than the Three Gorges Dam project, though reliable statistics are not readily available. Some estimates of displaced persons created by the Sanmenxia Dam cite 300,000 migrants. But the longterm effects on individuals, families, and communities have been perhaps even more severe. The project was designed without strong geological or hydrological expertise, and it has failed fundamentally to achieve the intended goals of the project. (Journalist Xie Chaoping was arrested in 2010 for writing a sober appraisal of the failures of Sanmenxia in The Great Relocation; link.)

French scholar Florence Padovani has studied the topic of involuntary relocation in China in detail. Her contribution to Forced Migration and Global Processes: A View from Forced Migration Studies provides quite a bit of factual background on the important hydraulic relocations. Here is Padovani's definition of forced migration:
Forced migration, as it is understood here, involves populations which are uprooted by development programs. These individuals are internally displaced; even when they move thousands of kilometers away from their birthplace, they remain inside national borders. (92)
And here is her summary description of the relocation of population for the SMD project:
The way resettlement was implemented in the SMD case is typical of the way the government at that time considered people to be displaced. The authorities, from top to bottom, adopted the same line of action in a unified rhetoric. As for the resettlers, they did manage to raise their voices through very traditional ways, such as petitioning the central government ..., but also through street demonstrations and even physical fights. ...  We can summarize resettlement policy during the revolutionary period by the following points: First, the vast majority of displaced people were rural dwellers, mainly farmers. Second, they were not involved in preliminary assessment projects and were informed of the government's plan at the last moment. Host communities were not involved in resettlement plans either, so some serious conflicts erupted when newcomers settled on their land. And third, displaced people were not well compensated and their relocation was not well planned, leading to their impoverishment. (97)
Padovani's ethnographic study of Three Gorges Dam emigrants in the Shanghai region sheds much light on the human consequences of forced migration in China (link). This research provides a report of interviews with a few dozen migrants in the Shanghai region.The great majority of displaced people are farmers. So their life prospects following a move are bleak. They need access to land, secure housing, and access to unskilled employment. And none of these  are in ready supply in most destinations. More intangibly, they need access to kinship networks and mutual aid societies of the kind that were available in their home communities. But forced migration has thoroughly shattered those sources of mutual aid. So the lives of these million-plus migrants have not been made more hospitable as a result of the project. 
The disintegration of the social network due to forced migration means that the 'domestic order" (local and personal ties), "civic order" (equality and solidarity), and "market order" (economic performance) have to be rebuilt. (107)
image: The Chinese Dust Bowl, The Walrus (link)

Yellow River communities have experienced another wave of resettlement in the past thirty years, this time along the upper reaches of the river in Gansu and Inner Mongolia, in response to the combined pressures of development and desertification. Irrigated farming communities have been uprooted multiple times as local authorities have sought to make use of scarce land and water resources. (Benoit Aquin's photo essay on this region in The Walrus is a great exposure of these dustbowl conditions.)

Reading the story of these massive dam projects and their state-sponsored programs of involuntary resettlement illustrates some of the perverse processes of development that James Scott describes in Seeing like a State: How Certain Schemes to Improve the Human Condition Have Failed. Scott doesn't consider modern Chinese examples, but he finds a raft of critical problems of environmental impact, social disruption, and human suffering associated with major water projects in other countries. His summary thoughts seem to apply to China's development experience as well:
I believe that many of the most tragic episodes of state development in the late nineteenth and twentieth centuries originate in a particularly pernicious combination of three elements. The first is the aspiration to the administrative ordering of nature and society, an aspiration that we have already seen at work in scientific forestry, but one raised to a far more comprehensive and ambitious level.... The second element is the unrestrained use of the power of the modern state as an instrument for achieving these designs. The third element is a weakened or prostrate civil society that lacks the capacity to resist these plans. (88-89)
(Here is an earlier post on Scott's critique of state-driven modernization projects; link.)

The Sanmenxia Dam is an unmistakeable failure; siltification and massive pollution have turned out to be irresolvable problems along the length of the middle and lower reaches of the Yellow River. Experts are now calling for the removal of the dam altogether. There are very strong indications that the TGD is likewise an environmental and social disaster, though this is still not completely resolved. And for the several million people who were forcibly relocated, the Three Gorges Dam has often turned out to be a catastrophic turning point in their personal lives.