THEORY AND MIGRATION Towards

The following article discusses the role of the human agency in migration theory. In connection to that special interest is focused on the structuration-theory where the human agent plays an important role. The relationship between the structuration-theory on the one hand and the concept of time-geography on the other hand is discussed. It is argued for that the development in time-geography towards micro-simulation approaches opens new doors for the debate concerning the relation environment – human agency. At the end the article argues for a combined approach where migration is understood from two viewpoints; the structuration theory and the time-geography.

Mots clés : théorie de la migration, théorie de la structuration, géographie du temps "Finding a general theory of migration with universal validity and applicability is the perpetual dream of those working on migration research. To the ambitious this has become an obsession; to the more realistic it has remained a fond hope" (Chang 1981, p. 305).

Introduction 1
In this article migration-theory and its use as a platform for classification is discussed. Furthermore, the problem related to the so-called structure -agency dilemma as well as the similarities between the structuration theory and timegeography is brought up. Finally, the article discusses how the phenomenon of migration can be understood and interpreted.
A point of departure for elucidating the distinction between different theories of migration is the differences between theories about societies with individuals on the one hand, and theories about individuals in societies on the other hand. The former relates to structural causes for migration, while the latter relates to theories where the individual reason for migrating dominates. The division between macro-and micro-studies can be seen as a parallel division. Even here migration patterns can be explained either in macro-terms, where individual behaviour is seen as a result of changes in the surrounding (structure), or in micro-terms where individual decision-making and individual values are given more substance. This division is not without objections, when macro and micro normally refers to the aggregate level of the empirical data. An alternative approach suggested by Westert, Verhoeff (eds 1997) taking into account both macro and micro levels in order to study human behaviour. That means that scholars can study macrophenomena and micro-processes in a contextually manner.
The selection of time-frame seems to play an important role when it comes to classification of migration theories. For example, looking from the perspective of a lifetime, movement can be interpreted as a result of individual free will. In a longer time-frame it may be reasonable to understand individual relocation as a result of long-term changes in social and economic structures. Therefore, migration theories ought to be adapted to different basic societal circumstances. There are sharp contrasts between the refugee problem in Burundi or East Timor and return migration in Sweden! There are also major differences between short distance migration and long distance migration. At first glance all migration is a question of human movements over the earth`s surface. The reason to migrate varies. What influence has the structure over the individual decision and what choices are open to the migrant? Do circumstances vary over time between the surrounding (structure) and the individual, and between development levels in society? Zelinsky 2 considered levels of development in his migration theory (Zelinsky 1971), but he does not take into consideration the relationship between environment and migrant. Do we migrate because of forced circumstances or do we migrate due to individual, voluntary, driving forces? Decision to Move from Origin  (Eichenbaum 1975 Eichenbaum (1975) classifies different types of migration from voluntary migration and forced migration (see figure 1). It is however a little problematic to define different types of migration in this manner, while migration is rarely that simple to categorize.
Roger Andersson (1987) conducts an intellectual experiment in his study supposing that all movements are forced, which means that the individual forfeits his right to decide by himself. Even if many human actions are influenced or even determined by structural circumstances, interference in the individual`s right to decide is an infringement. Political refugees from Chile, the compulsory transfer of Baltic people during World War II or hidden minorities are all subjected to, more or less, forced movements. It has to be said here that structural causes do not have to be inflicted. Despite the determinant power of structures, structures can also be seen as a freedom to use the possibilities that originate within the structural framework.
Despite the demarcation-line between theories about societies with individuals and theories about individuals in societies there is no dramatic lack of integrative migration-studies. However such studies, where complexity and important factors are related to each other, are less numerous than their opposite. My view of integrative studies is that they contribute valuably to the understanding of the context that surrounds the migration-decision making process. During the last decade several studies have been carried out, with greater emphasis put on individual decision-making and the decision-making process (see for example Andersson 1987, Löfgren 1990. The time-geographical perspective has also contributed by creating better opportunities to study and analyse the relation individual -environment (Holm, Mäkilä & Öberg 1989, Åqvist 1992.
From a macro-orientated perspective it could be claimed that, within modern societies like Sweden, factors like labour market, housing-market, careers, education etc., are of great importance when explaining migration-patterns. In this sense it might also be possible to regard social networks from a macroperspective. On the other hand, this unique individual behaviour constitutes an important ground for explaining migration.
During the 1980s a combined approach was developed, where a supposed interplay between individual and surrounding or between actor and structure was seen as vital for explanation (Giddens, 1979(Giddens, , 1981Gregory & Urry, 1985;Åqvist, 1992). This approach is known as "structuration theory". In the following the structuration theory in its relation to time-geography will be discussed. At the end of the paper I will come back to the importance of these two approaches in the field of migration theories.

The Structuration Theory
During the last fifteen years the increasingly criticised traditional parts of the structural school combined with a revived interest in different forms of individual oriented theories has resulted in several attempts for new, alternative approaches within social science.
It is particularily the relation between individual and social structure that has, for the most part been in focus. Giddens (1979Giddens ( , 1984 develops an alternative approach within the structural school in the "theory of structuration". In brief, he argues that the human agent or the human factor must be given more attention. In Giddens model the human agent plays the principal part in a social system. The subject or the human agent is considered not to have perfect knowledge about the society or reality we live in 3 . The point of departure in the model is that rationality lies behind every given individual´s behaviour. This rationality is based on the individual`s knowledge and information about society and the empirical world.
Humans as subjects are not only isolated singularities but also interpreters of the world in interacting social systems. With this, Giddens means that changes or social actions take place within a framework, a kind of empirical structural network. The most important element in the frame is place 4 .
Irrespective of whether a place refers to a town, a cultivated area or an industrialised complex, the place always reflects as a consequence of human activity. The place as a phenomenon concerns requisition and transformation of place. This is linked to reproduction and transformation through time and space. Pred (1984) puts more emphasis on the physical surrounding and its influence on the social structure, than Giddens. Place is just not a thing on the earth`s surface, a framework for activities and social interactions. The time perspective has to be added. The exploitation of natural resources and the transformation of land are obvious signs which speak in favour of the need for a time-perspective.
Concerning the blurred social structure Carlstein (1982) identifies a problem. The idea of the structure as a code independent of time and space does not fit in the structuration theory. Giddens means that each structure is unique for each moment, the code or structure changes in the next moment. "In the latter, each structure is unique to each moment, and alters as that moment is replaced by a new one. Hence, the output of two subsequent moments can never be the same, since life passes in transformation with time" (Carlstein, 1982, s. 48) 5 . This is not a revolutionary discovery. The maudlin Greek philosopher Herakleitos (ca 540-480 f Kr) meant in this manner that everything is floating (panta rei). His line of thinking maintained that existence undergoes an eternal change and is constantly in movement. Even Pythagoras worked with similar thoughts. "Everything is changing, but nothing perishes".
In Giddens theories two main lines of thought can be identified: The first is "the theory of structuration", the second line of thought is to investigate and explain the main currents in modern societies. Gidden`s point of departure is the conflict between "action" and "structure". On the one hand, there are scientists that argue that society or the social structure is a precondition for the individual and the 4 The concept of place creates confusion in the debate. The time-geographical Swedish concept "rum" has no direct meaning or good translation in english. The concept of "rum" comprises both geographical room and social room. Giddens uses the word "locale", where the social room is depending on the geographical room. Johnston (1986) translates the word "locale" to "space". Giddens defines "locale" as "A physical region involved as part of the setting of interaction, having definite bounderies which help to concentrate interaction in one way or another" (Giddens 1984). 5 A question of more philosphical charactar is the importance of memory. How is the code or structure influenced by the passed? Does the code have a memory? A trajectory shows a frozen moment, a development over time with a time-memory. Giddens discusses however not in terms of structurememory, but in terms of constantly changing structures. The actual existing structure interacts with the individual, who interprets and acts and by the human acts the structure changes. This must also mean that the structure is built upon collected experiences and thereby one could say that the structure has a time-memory. purpose of structure is to delimit individual acts. It is this structural delimitation which states the real objects for analyses (Durkheim -Levi-Strauss -Althusser).
On the other hand, this traditional approach has been challenged by scientists who argue for the viewpoint that social science has to take into consideration that social surroundings also consist of individuals` own interpretations of the structure. This trend was advocated, for example, by Max Weber and by scientists influenced by hermeneutics, phenomenology etc (Held & Thompson, 1989). The sociologist Zygmunt Bauman (1987) describes Giddens theory as a theory with two aims, a) to dethrone the concepts of structures as a determining external factor for action and b) to deny the existence of randomness in human behaviour. Furthermore Bauman claims that Giddens redefines the concepts of structure by arguing for the existence of some kind of a mysterious force or power. Giddens reifies social patterns as processes or things. One consequence of this is that certain events are more likely to occur than others are. Certain actions increase the possibility that the consequences of given actions will follow a certain pattern when taking the original action into consideration 6 .
Giddens theory is an attempt to think right through action and structure and to describe how these are woven together in daily life. The difference between structure and system is obvious for Giddens. The former relates to rules, restrictions and resources and the latter to the pattern of social relations in time and space. That implies that the social network can be characterised as a system which is influenced by, and influences, the structure. Within the system-analysis approach one also differs between system and structure in a similar way. The structure is more rigid and contains links and nodes, while the system is something going on between the nodes via the links, usually called flows. These flows are more variable to its nature.
Like the rules of grammar, structure is both "enabling" and "constraining" it enables us to act as well as delimiting the courses of possible action. By focusing on the generative character of rules and resources, we can see that structure is both constitutive of everyday action and, at the same time, reproduced by that action -a phenomenon that Giddens refers to as the "duality of structure" (Held & Thompson, 1989, p. 4).
The structure-concept within the structuration-theory is also criticised. The critics are summarised in Åquist (1992). The main critique is that structures are always understood (interpreted) as frameworks for action and are not seen as forced powers. Hereby the concept of structure in the structuration-theory diverges from the structure-concept in Marxist-theory, which advocates a more determined view on structures.

Structure and Realism
Giddens theory is criticised in many disciplines, but especially by sociologists and geographers. One of the key-issues in this critique is whether his theory originates from realism or from structuralism. The interpretation of the structuration-theory and the importance of the concept of realism have become an arena for many questions and discussions.
The realistic worldview implies that the exterior world exists independently of mankind and his consciousness. In social geography this is made apparent in different ways. Marxism is materialistic in the sense that the material comes first.
The emotional life depends on the material. No emotional experiences can exist without the physiological processes in the brain. The debate about the concept of realism is wide and in some senses a little bit confusing. David Harvey claims that it is ridiculous to abandon the historical-materialistic view. The attitude of rejecting the idea that individuals create their own history, both as individuals and as a collective, under material conditions and that the individuals are not responsible for their acts, to favour concepts like structure and agency, is politically meaningless (Harvey, 1990). Massey (1984) and Sayer (1984) have developed an interest in places, events and process singularities. Sayer is critical of how Marxian theory defines individuals and single events. He argues for a more realistic philosophy, which combines a free play for coincidences with an understanding for general processes.
What then is realism? Sayer (1984) describes realism and positivism as two contrasting methodological principles. Sayer compares realism with the traditional conception of positivism and empiricism. An approach to research in social science that is built on realism is still orientated towards empirical methods. This may not be mixed up with empirical research. The difference between the two approaches is the view on acquiring knowledge. Empirical knowledge as such is non-problematic. The use and interpretation of knowledge produces this whole difference. Realism represents an alternative method for empirical research (Asheim, 1990). The main purpose of empirical research with a realist-orientated approach is to explain the observed events by postulating mechanisms that produce these events (Sayer, 1984).
According to Sayer, empirically observed events are not just a result of direct causal connection in and between structures and mechanisms. On the contrary, events can also be seen as a result of historical and territorial conditions. The history and the importance of places are given more elbowroom. In the discussion about the relation between agent and structure it is implied that science (or the scientist) must try harder to understand what the actor understands or what the actors intentions are determined by in his/her actions. We can not just use empirical observations when we want to understand individual behaviour.
Besides observations and causal explanation-models the question "why?" should be raised at this juncture. Why do people act the way they do?. Besides a world which we observe and formulate, there is a world that exists independently of our consciousnes. There is a world we are not aware of. It is attempting to draw a parallel to gods-faith and metaphysics, but this parallel would be misleading. Realism should rather be considered as a philosophical analogy to subjectivity and objectivity. The subjective is the world as it is interpreted and reformulated by the individual`s conscious. The objective world is that part of the world that exists independent of the individual`s conscious. Johansson and Liedman (1993) write that knowledge-realism appears to be the superior perspective. Realism means simply, that there exists a reality independently of humans and that humans can get in touch with this reality through praxis (Johansson & Liedman, 1993, p. 242).

Time, Space -Time-geography
Within the concept of Hägerstrand´s time-geography, time is considered as a contingent flow and space as a finite resource. The interplay between institutions and new and old forms of organisations for human activities in the time-space perspective has become a central theme in time-geographical research. The geographer looks on every single place as a result of general rules and common social structures, which are applied in different regions. The place is also considered to creates unique combinations together with the actor`s interpretations of general rules and common structures.
Time-geography is sometimes seen as a time-spatial theory in the discipline of human geography, but time-geography can also be seen as an approach where hypotheses can be tested and theories developed (se for example Holm, Mäkilä & Öberg, 1989). Time-geography has a wide ranging capability and use. Commonly it is claimed that everything is connected to everything else and that timegeography aims to describe events in time and space. The basic theories behind system-analyses are partly built on this view. Time-geography could in fact be seen as a developed system-analysis. In time-geography it is the physical phenomenon in the first case that determines the constraints or conditions for actors in the real world. By applying the structural perspective in time-geography it will become possible to see the smaller picture in order to understand the larger, allowing to investigate complexity on a small scale. Hägerstrand emphasised the importance of delimited activities (pockets of local order). In the field of structuralism these pockets of local order are seen as parts of larger systems. The concept of "regionalisation" is closely related to these two approaches. These approaches criticise firstly the conception concerning the division between the micro and the macro-level within science. Secondly, the conception of welldefined societal limits is questioned. To put it another way, one could claim that there are no isolated systems or single levels, they all depend on every other system and/or level.
Giddens emphasises that the time-spatial relation is of vital importance in social theory since time and space are basic bricks in social systems (Giddens, 1981).
Giddens uses time-geography as a description-system where the characteristic forms for time-spatial relations can be viewed. But time-geography is at the same time something more than a description-system for time-spatial movements. It visualises patterns of movements, which have their basis both in physical conditions and constraints as well as in social relations and constraints. On the one hand individuals and events and their time-spatial connections are visualised and on the other hand the importance of the constraints for the time-spatial connections are obvious. With a time-geographical approach we will be able to study, for example, the time-spatial consequences of social relations. The perceptible is in fact the object in time-geography. The graphical notation is more important than the conceptualisation (Åquist, 1992). In this matter Giddens, among others, has a divergent opinion. Giddens stops at time-geography as a description-system for the geographical outcome of social relations.
Here it should be noticed that time-geography emanates from Hägerstrand`s search for a way, or a method, to represent human actions (Hägerstrand, 1972). The result was the notation, or in Martin Gren`s words, "the musical notation" (Gren, 1994). It could be argued that notation is just another aspect of spatial science, however it does not only visualise the physical perceptible phenomenon, but also the spatial consequences of, for example, social processes.
Critics of time-geography argue that it is not enough to study the perceptible. We also need an explanation for the perceptible. Social theory needs to be added to time-geography. Several researchers have mentioned this approach, but few have practised it. As Gren (1994) points out, this was not Hägerstrand`s intention from the beginning, but, on the other hand, this is not an obstacle to trying to incorporate social theory in the concept of time-geography. The criticism has in fact two points of departure. Firstly, there is a lack of an insider-perspective, which gives individual values with an increasing importance (Åquist 1992). Secondly, social conditions should be more emphasised in order to understand and explain events and processes in the time-geographical notation. Hägerstrand himself writes in the essay "Tidsgeografi" (1991) that man differs from animals in the sense that man plans in advance and this type of long-term planning aims to reduce the random in future events (Hägerstrand 1991). This human long-term planning consists not only of knowledge of the physical environment, but also of an idea of the importance of social relations. To reduce time-geography in this situation to physical reductionism is maybe to go too far. Within the framework for a "project", sequences of events can be described. Material and physical conditions can primarily subordinate an event, but the event in itself contributes to the creation of new conditions for the next sequence of events and so on. In this way, social conditions, values and meanings can be represented.
One attempt to bring time-geography forward by combining model and empirical data was made in the so-called HÖMSKE-model (Holm, Mäkilä and Öberg, 1989). The study, and the model, can at first sight seem to be a micro-simulation model, where the rules are governed by statistical probabilities based on empirical material. This is, however, not the whole truth. In the model attempts were made to create rules, rules that affect individual`s behaviour. These rules were divided into three categories; what an individual ought to do, can do and wants to do in a given situation. The design of these rules is partly built upon theories about social relations, cultural values, norms etc and how these govern (or do not govern) our lives. A model is always a simplification of reality linked with the researcher`s values, the moral of our own time and physical constraints. But a model can also constitute a valuable contribution to our search for a better overview, explanation and understanding of the world. As I see it the HÖMSKE-model is a model where social theory and empirical data contributes to the construction of rules in the field of micro-simulation 7 .
The criticisms put forward against time-spatial relations are, then, based on whether the intersections in this relation are acceptable conditions for constituting social structures. Spatial structure is not comparable with spatial interaction. Giddens (1979) talks about three aspects of structure "signification", "domination" and "legitimisation". All social interaction means connection between time and space in the sense that there is always a carrier of social interaction. This carrier, or intermediator, is needed to overcome the gaps in time and space. These gaps consist of what we usually call distance. In societies where personal contacts dominate, the mediating vehicle consists of physical institutions. Other communication media (such as letter, telephone etc) can overcome larger distances in time and space (Giddens, 1979;Carlstein, 1982).
The determined human agent in time-geography is a target for critical remarks. Time-geography uses the agent in a naive way. It is what is going on inside 7 Simulation is both a reproduction-technique and a problem-solving process (Holm et al 1989). The observed reality is reproduced as a computer-program. When single individuals are reproduced in a simulation-system it is called a "micro-simulation". Just as in reality the model can reproduce randomized elements, stochastic simulation: The arguments for using micro-simulation are mainly related to the difficulties to use empirical results in macro-models (Orcutt, 1976). Social conditions and relations between actors are both richly varied and complex, and consequently these systemrelations are hard to summarize in only one variable. The purpose of micro-simulation models has changed over time. Holm, Mäkilä, Öberg (1989) do not aim to explain the whole world and make secure forecasts about the future. The simulation-model is a summary of knowledge. The ambition level in the field of micro-simulation has, according to Holm et al (1989), been moderated during the last few years. Today (1989) its purpose is to understand some parts of the mechanisms in development that can be influenced (ibid 1989). Scientists working with the micro-simulation are well aware of the critizism. Holm, Mäkilä and Öberg compare the micro-simulation model with a piece of art, which should be judged by our age-time and also, at the same time, needs to be constructivaly criticized. The main effort of micro-simulation is its strong relation to theoretical assumptions. Theory and empircal data have both strong implications in micro-simulation methods.
individuals, in their consciousness, which is of interest. A time-geographical description can answer questions like "where" and "when", but not "why" (Gren, 1994). These statements are rather typical for the criticism of time-geography. Gren (1994) writes that a possible reply to the criticism, and the difficulties of adopting the criticism is maybe the original attempt to note sequences with timespatial constraints (musical notation). The time-spatial restrictions (constraints) are often rather physical, which means that"...it is the step towards the natural sciences, and not to social theory, which is the logical consequence of Hägerstrands time-geographic approach" (Gren, 1994, p. 86).

Concluding remarks
"For positivists, the goal is social engineering, manipulating society towards certain goals and away from others; for humanistic scholars, it is increased self-and mutual awareness; and for structuralists (especially those linked to critical theory), it is emancipation, the removal of the false ideology that clouds interpretations of the mechanisms that drive society" (Johnston, 1986, s. 156).
It is inevitable for a social scientist to choose an attitude. Johnston (1986) writes however that with the guidance of Gidden`s structuration-theory is it possible for social scientists to bring together an empirical based method with humanistic approaches, which emphasise understanding and importance of individual actors. The structure changes all the time through the myriads of actions, which continuously occur. Giddens et al (1982) touches especially upon two aspects of individual behaviour: The ability or the right to choose another alternative and the individual's capability of acquiring knowledge, everything that the individual in the society knows about the society and the conditions for the individual behaviour in society. The scientist needs to have a deeper knowledge of the underlying structure and the observable, empirical world, but most of all he/she must have an ability to interpret the individual behaviour in the sphere between structure and the individual.
It is possible to argue that Giddens got the idea of the structural skeleton from structuralism and thereafter added one more interpretation aspect. The researcher interprets the result of the individual interpretations of the structure. Giddens calls this double interpretation for double hermeneutics. Giddens main principal concerning the structural skeleton as mediation for and the result of social acting is, in a way rather simple and without objections. The other point of departure is Gidden`s use of the time-spatial dimension.
The sociologist Thomas Brante from Gothenburg has criticised some parts of the structuration theory (Brante 1986). Among other things Brante´s opinion is that Gidden`s theory of structuration is somewhat watered down. There can hardly be any social scientists, who believe that social structures exist without human actions, not even the most enthusiastic structuralists (Brante 1986). The critics applying time-geography are also mentioned by Brante, however in its most trivial form. A superficial observer may criticise the description system of timegeography as being natural, namely the fact that we are doing different things at different places during different times.
The central constituent in the structuration theory is the interactive dualism between agent and structure and the opinion that structures are both the medium for,. and the result of, social actions. Brante (1986) questions if it is possible at all to unite agent and structure. Giddens intentions could also be applied in a functional analysis.
It is impossible to explain and understand the world in one general equation and this is not Giddens and Hägerstrand`s intentions. The creation of a socialtheoretical framework for social actions, a physical contextual-and a connective framework for human actions, are both important steps towards a wider contextual framework for the explanation and understanding of human behaviour. It is therefore a pity to claim that the structuration theory is a theory and this could also be said about time-geography. They should not be regarded as theories. It is all about approaches or views.
A combination of methods and approaches will open new doors and create a better understanding, for example, for the migration process. In order to better understand migration behaviour and in order to make better forecasts a structuration framework in a time-spatial context is indeed needed. In this context some interesting contributions have been made (see Hammar et al 1997, Stjernström 1998.
Structuration theory and the time-geographical approach gives two dimensions that can function like an understanding framework when working with migration. It is in many ways quite simple. Time-place and space matters. Events takes place and time-space constellation gives constraints to actions. The relations between time and space is changing over time and so does the social landscape. New conditions are created constantly, conditions that affects and governs the individuals actions. By studying migration patterns and migration behaviour these changes and the importance of time, spatial and social constraints affects the individual behaviour.
So, what about the system theory and the phenomenon of migration? It is reasonable to discuss in terms of the time-space framework. It is always there. Actions take place on different places and between different places over time. The relation between time and space is not static. The relation could be described as a plastic relation or in other words "plastic space". That means that time and space interacts continuously. This is essential because the time-spatial conditions affect the individual possibilities to act.
The description and understanding of the phenomenon of migration is heavily related to the level of analysis. Human movements across the surface could be described as it is, as physical movements. Human movements could be understood as a reflect of physical conditions in the environment or it could, at last, be understood as a result of social conditions in the society. The general problem with migration theory is its focus on a special topic or condition, for instance theories related to distance or social conditions. It is harder to find or discuss in terms of a more holistic view on migration. This could be seen as a shortcoming but also as a challenge for future research.