Deze samenvatting is gebaseerd op het studiejaar 2013-2014.

Chapter 1. The analytical tradition in Sociology

Dissecting the Social: On the principles of Analytical Sociology book will focus on the approach referred to ¨Analytical Sociology¨. Leaving behind the empiricism based on sociological variables as well as all the old writings from big social theorists.

What is analytical sociology? Analytical sociology tries to explain the complexity present in any social process. This means dividing the process into pieces -¨dissecting¨- and afterwards bringing the most important parts together.

Analytical Sociology approach is divided in four basic elements that need to be overcome: explanation, dissection and abstraction, precision and clarity and finally, action will be taken on the social phenomena.

Explanation

An explanation will give an answer to questions like, why something occurs, why things do change over a period of time, etc. This book uses the mechanism-based approach in order to find a clear explanation to the social events. To summarize, this approach will focus on a specific sample of social processes, activities, and actions taken by its actors After that, all of them will be linked up to each other. This way, we will be able to find out what kind of social phenomena we want to explain.

Dissection and abstraction

These two words are key elements in the analytical approach.

The dissection process will allow us to have a better image of every single entity and about all the relevant components that complete the process as a whole.

The abstraction process takes part when we individually focus on something that apparently is very important for us in order to solve the social problem. However, by doing this, we are moving out of the real focus on other elements that might be of less relevance.

Precision and clarity

A theory proposal’s goal should always have clear concepts and explanations that build it. This is the reason why a theory without any clarity and clearly understandable elements, may not help to achieve the goal. Moreover, clarity in the sense of precision on how we come up with a clear theory proposal is very important as well. In other words, we must be able to detect what are the differences that make our concepts and theories different from each other. Otherwise, we would not be able to compare them, to explain them, to observe them.

Action

The actors are the core element in the whole dissecting and abstracting process.

The society moves thanks to the actions of the actors. If there would not be actions undertaken by the actors, the whole social phenomena would stop.

It is said that in order to have a better understanding of why actors do what they do, we should analyse their behaviour. Furthermore, we must understand that sociology focuses on a range of actions taken not by one single individual but by the collectively sum of several individuals. Thus, collective member characteristics can then be catalogue as a social property.

Chapter 2. Social Mechanisms and Explanatory Theory

The importance given to social mechanisms in the analytical approach is because the social mechanisms bring about the final outcome of social phenomena. In other words, a set of well-defined activities, which are linked one to another will explain us the social phenomenon. This chapter two will basically focus on explaining the relevant framework that forms Sociology. To do this, the author will cover some important philosophical literature. To give this chapter a starting point, the different varieties of sociological theorizing. For this purpose a concept to what a theory is from a philosophical and sociological point of view needs to be given. Sociologists seen theory as the transmitting tool that society uses in order to express their feelings and problems. From a philosophical point of view, a theory is something abstract that gives sense to our knowledge in order for us to find an answer to explanations of events and processes. But what an explanation is exactly?

1. An explanation is not a description. Because a description gives us information for instance on how something changes over time and not why something changes. The difference between an explanation and a description is quite obvious. Even though, both concepts -¨interrelated, provide answers to different types of questions¨.

2.Explanations differ from typologies and taxonomies. We can say that typologies are tools that allow us to tag different social phenomena with a name. However typologies do not explain why we observe this social phenomenon in the way we do it. It is not possible to have an explanatory typology.

Moreover, it is relevant to mention one more terminological distinction between the philosophers of science and sociologists/other scientists, the former considers that causes and effects must be events. While the latter states that social states and individual characteristics as causes and effects.

3. Explanations provide plausible causal accounts for why events occur in a period of time and space. From here we see how there are three types of explanations. These are the following: Covering-law explanations, Statistical explanations and Mechanism explanations.

A) Covering-law explanations

Covering-law explanations- are theory-based, taking past theories or laws in order to explain concrete phenomena and using a deductive argument to come up with a general outcome. When we have an event, call it ¨x¨. We will try to come up with an explanatory answer for this event. In order to do so, we must assume that ¨x¨ does happen under certain general law. There is one disadvantage to the covering-law model; this does not describe well the characteristics of what is to be considered acceptable scientific explanations. According to the book’s author, the covering law theory has not been conducive enough to the explanatory theory. Thus, the problem involves lacking of restrictions on the content of the propositions. We can conclude that this model is not the sufficiently good for sociology.

B) Statistical explanations

The statistical explanations form part of a very practical model. Thus, it is of relevant importance for empirical research in sociology. The statistical explanations are way more inductive comparing to the Covering-law explanations. The methodology used in this model is to dissect a particular group/population into categories. This will be done though designing hypothesis. After that the hypothesis will be tested. From there on, an explanation to the proposed theory will be found. This builds a useful statistical model with a proper strategy to achieve. An important fact to bear in account is that a statistical analysis is simply a test of an explanation, but not the general explanation.

C) Mechanism explanations

The mechanism explanations are a core part of the analytical approach. The conceptual idea of what a mechanism is and has been described by different philosophers and social scientists like entities, causal mechanisms, social mechanisms, models, etc. However, Hedstrom-Swedberg approach and Machamer, Darden and Craver (2000) are the most satisfactory ones. Furthermore, we can say that mechanisms are formed by 1) entities and 2)properties activities that entities get involved with. These activities bring about a change; the type of change depends on the properties of these entities and how they are linked to each other.

But what is a social mechanism? A social mechanism describes a whole made of entities and activities that are organized in a particular way to bring certain type of outcome. These outcomes can be inequalities, typical behaviours of individuals given in different social situations and social norms. But what does define the structure of interaction? The structure of interaction is defined through the way in which actors (the core entities) are liked to each other. Following the structure of interaction will influence the final and type of outcome. In addition, the importance to specify mechanism has two reasons: 1) we try to reduce theoretical fragmentation to bring out structural similarities between apparently different processes. 2) To know more precisely the causal relationship between a cause and its effect. To be sure there is not just a simple correlation between them. 3) To reach more precise and intelligible explanations.

Differences and Similarities

After discussed the three most important traditions, now we will focus on its differences and similarities. As the books highlights mechanism-based explanations hold a narrower relationship with covering-law explanations, which refer to causal factors, than to statistical explanations. On the other side, statistical explanations is formed by factors that try to show where the difference on the matter to be studied relies on. Thus, statistical explanations can be seen as theoretical deductive arguments. Another difference between the covering-law explanations and the mechanism-based explanations lie in the fact that, the former’s explanations can be seen as perfectly general and without exceptions, while the latter’s explanations are the contrary. Mechanism-based explanations are more action-based, opposite to the covering-law ones. Finally, one of the differences between statistical explanations and mechanism-based explanations lie on the central issue of randomness and the stochastic nature of the social processes. But, we should not forget that social phenomena are the final outcome of various and different causal processes working together at the same time. Thus these processes may influence and overdo another process. That is the reason why we should take into account the relevance of treating most of the statements from an empirical perspective and not only about actualities. This is a well-known John Stuart Mill theory on mechanism-based explanations.

In conclusion, we have seen that when theories are explanatory, must contain questions on how things happen, why things do change over a period of time or why situations change in time and space. We distinguished between three types of explanations. The mechanism-based theory is the most suitable for sociological theory. Moreover, we saw how a social mechanism is formed by a group of events that all ordered together lead to a certain result, a social phenomenon. Thus, we can say that a social phenomenon is explained by the social mechanism.

Chapter 3. Action Interaction

In the previous chapter, we dealt with covering-law explanations, mechanism-based explanations and a statistical approach, which tells the relationship between variables describing different entities. According to Popper, we are looking for mechanisms that explain the causes and consequences of individual actions. Since these actions are responsible to form the social. Moreover, it is seen how the mechanism approach and the Weberian theory shared something in common. This similarity comes from not accepting aggregate correlations as explanatory until they have been driven by individual action. Thus, theories of action and interaction are the background for sociological theories. In order to find a proper type of action theory, three facts need to happen:

1. psychologically and sociologically plausible: The mechanisms are work must be explained in detail and the structure of social interaction should be taken into account as well.

2. Be simple: to grasp what the problem at hand is. There may be complex processes within the theory and that is why we should try to keep it as clear as possible. The sociological theories have three types of components as follows;

• An individual-action component

• A component describing the structure of interaction

• A component linking micro-actions to macro-outcomes

3. Action should be explained in an intended and clear way: an action should be explained making reference to the future that we are trying to bring about in.

A possible example that could fit these three types of sociological theory is the Homo economicus. However, it lacks psychological and sociological realism.

Before focusing on how the action of a single actor can influence the actions of fellow actors, the action theory DBO, we should mention three types of action theories:

1. As an interpretive tool – understanding the behaviour of one single actor

2. As a predictive tool – to anticipate the behaviour of an determined actor

3. As a theoretical mechanism of a sociological theory

The difference between these three types of action theory lies on the characteristics of the actor. For the first two types, the theory focuses on a real behaviour of existing actors. For the third one, the theory focuses on abstract ideal typical actors.

The DBO theory is based on the following terms D (desires), B (beliefs), O (opportunities) for the analysis of action and interaction. The term ¨action¨ stands for what individuals do intentionally. The term belief can be seen as a proposition about the world held to be true. A synonym for desire may be the verbs to wish or to want. Opportunities stand for designing a set of action alternatives that are at disposition of the actor apart from his beliefs. But how does an action occur? The cause of an action is grouped by desires, beliefs and opportunities. The way these terms are linked to each other makes the actions to differ.

Three ideal-typical explanations relating to causal efficacy based on beliefs, desires and opportunities are the following:

1. Belief –based explanation

Through the behaviour and actions of the social, individuals influence the beliefs and proceeding actions of others.

According to Merton, a false belief provokes a behaviour that soon or later will make the false belief to be true. An example to this, bankruptcy of a bank provoked by starting rumours of a possible insolvency, thus depositors will start withdrawing their money, just in case and to be sure. This will feed the rumour, leading to a social belief the insolvency may be true, further withdrawals come to effect, the belief will be strengthen by the actors. This belief-mediated interaction mechanism where a single individual´s belief has a value on the performing act of a group of individuals, follows a logic of imitation. But what does happen if the beliefs about other´s beliefs need to be coordinated? We should try to explain the belief-action history in order to establish an understanding why individuals act and do what they do. Furthermore, another theory that explains, how the mechanisms of a non-intentional type work out when the beliefs of an individual are influenced by the actions of other individuals, is the cognitive dissonance theory. The cognitive- dissonance theory stands for psychological discomfort happening when an individual is exposed to two psychological inconsistent cognitions, these can be beliefs, desires, attitudes, etc. In addition, two cognitive elements can be considered dissonant when the importance an individual gives to elements is related to how these elements are being observed. Finally, dissonance´s magnitude is a function of the strength in which a pressure is able to reduce the dissonance.

2. Desired-based explanation

First of all, we should make a clear distinction between primary and secondary desires. After that, we can divide three types of desire-mediated interactions.

1. When others do an action (Z), this action influences how much I desire to do the same action (Z). So, the individual´s action is dependent on what the other´s action is.

2. When others do an action, the individual desires to do or not do their same action. The individual has a primary desire to act like others. This may be seen as sign of conformity.

3. When others do an action (Z), the individual believes that doing the same action increases or decreases the chances to choose another desired action, let´s say (A). The individual has a secondary desire to act like others.

For all these cases, the mechanism changes but the outcomes are the same; all the individual´s actions are influenced by the other actor’s choice. In:

1. the actions of others are the cause of the individual´s desires.

2. the individual actor acts according to what the other´s did previously.

3. Opportunity-based explanation

Finally the last type of social interaction, the opportunity-based explanation is characterised when the action of an individual actor does influence the action of another actor and does influence his opportunities as well. Examples are, criminal acts, differences in social mobility rates (job mobility), political opportunity structures, etc. The most important thing here is that opportunity interactions create social inter-dependencies, which can be of extremely explanatory importance.

Later, as result of interconnections between the DBO theory, we find three important factors:

1. Adaptive preferences: where actors only desire what they do know they could get. There is a causal interconnection from beliefs to desires.

2. Counteradaptive preferences: are causal connections from beliefs to desires as well. Here actors only desire what they believe they cannot get.

3. Wishful thinking: are causal connections from desires to beliefs in which actors believe just what they do desire to be the case.

Besides the DBO theory, we may turn for a change by using rational-choice theory and learning theories. The rational choice is very useful, however for some specific situations it does not have the power of an explanatory story. This is the reason why it will not be the first preference on sociological theory. For learning theories, actions are decided based on past actions of other actors. It is a backward-looking theory. Since the DBO is a looking forward theory. The learning theories are not an alternative to the DBO. By the moment, the DBO theory is the most preferable compared to any others.

In conclusion, the DBO theory comes to play when studying how social situations where actors influence their desires, beliefs and opportunities. And how a collectivity formed by these actors bring about intended and unintended outcomes, which are based on desires, beliefs and opportunities.

Now let´s see the case when a group of actors act in a similar way but do not have anything to do with social interaction. In other words, acting by being influence on one another´s beliefs and desires. There are three types of processes: first one it is an environmental effect, where the outcomes of this process may be mistaken for a social interaction´s outcome, social interaction. The second process, it is a social interaction, where the actions of a group of actors influence the individual´s actions. Finally, the last process is the selection effect, where the actors decide to take a particular action without reason.

To end up, we will deal with the different types of mechanisms.

Chapter 4. Social Interaction and Social Change

The focus here is on the macro-level or social phenomena where the actions and behaviours of individuals are brought about. The types of social phenomena are collective properties, which cannot be described by any single collectivity member. Examples are: Typical actions, beliefs or desires, Distributions and aggregate patterns, Topologies of networks, Informal rules or social norms. We have seen how several authors have explained social phenomena. Durkheim main explanatory factor was the extent of social cohesion. For Weber, was the existence of religious norms and for Coleman´s was the topology of a network. All these social phenomena had an explanan and an explanandum. However, sometimes it is not enough to assume that one social phenomenon causes another neither to assume there is always a correlation between the cause and its effect. We should always be able to open the ¨black box¨ . The black box will allows us to see beyond the social mechanisms. To see how actors influence their actions, and how the actions bring about a social phenomena that looks for an explanation.

The ontological and methodological point of view

As we mentioned before, the mechanism approach implies that interaction between actors bring social outcomes, these outcomes will later influence further actions. So this approach tries to explain how the social and the individual mutually influence each other over time. Further and on the one hand we can count with other theory described by Archer (1995) who has an ontological view of the social world as stratified. This means the structure and agency are ¨neither co-extensive nor co-variant through time¨ Why? Because, both have separate emergent properties, which are able to change, thus be independent of each other in time. On the other hand, Brante (2001) as Archer, he also believes reality is stratified, thus there is not ¨non-reductive causal mechanisms¨ operating at each different level. So, when social reality is ontologically divided and the social has the causal powers that realists attribute to it, so then their conclusions would be assured. This would lead to a separation of the individual and the social- division of labour would make sense.

The relation between the ontological status of the social and causal power comes from agreeing that individuals exist and have causal powers that allow them to make a change and go beyond social expectations. What about the realists? They do think this statement is true for society and structure. Why? Because society cannot be observed as such, thus a perceptual criterion of ontological existence cannot be used. In order to identified this ontological existence we need a good method in which causal effects could be seen.

However, this is not possible and this is why realists have not convincing theories about the causal efficiency of social entities. Despite that ontological collectivism does not offer much to sociology, the collectivist position on methodology, more precisely on the idea of ¨ supervenience¨ comes to play. In a principle, the concept of supervnience was used to characterise the relationship between the physical and the mental (Jackson, 1996). Now, we will use it to describe the relationship between the individual and the social. With the letter ¨S¨ standing for social property and the letter

¨I¨ standing for individual-level properties. If the individual-level properties of two groups are the same, then their social properties would be the same as well. This forward leads to assume that two groups having differences in their social properties will also differ in the individual properties. However, it cannot be assume that two groups with identical social property (S) will have exactly the same individual-level properties (I). Why not? Because, identical social properties can be displayed in many different ways. In addition to this, there is another point of view brought about by Sawyer. Where one specific type of social property is always followed by another type of social property, meaning that without any similar existing law-like relations at the individual level. With these conditions, Sawyer affirms that the only position to be considered would be an ontological individualist and methodological collectivist. However, there is not data to confirm this theory so we will leave it behind for the moment. Concluding, we can say that the supervenience theory presents a social level that is not ontologically individualistic but at the same time depends on individual –level properties. Besides Sawyer´s view, we can see level distinctions, a methodological distinction between different ¨mechanism levels¨ , meaning that individual actions would be at a lower level than the social phenomena is and this social phenomena be at the same time, be lower than actions or beliefs. Finally, it would be easier to see social phenomena from an epistemological view, this refers to see social properties which cannot be known by anticipating which emergent properties formed part of the phenomena before.

The individual and the social

In order to understand the actions of individuals, which led to complex emergent social phenomena, the interactions and interrelated parts forming those actions need to be studied deeply. The fact that changes in the way individuals are related to each other can lead to consequences for the social outcomes. The micro-level in which individuals interact with each other brings about different outcomes (social phenomena).

This social phenomena is sometimes very complex, so that is why we use formal analytical tools to solve it. Summarizing, social interactions are relevant in order to explain and understand certain social phenomena.

As conclusion to this chapter, two facts negative facts on the support given to the critical realists, first they did not have a convincing theory to support the causal autonomy and causal power of social entities. Second, they do not put too much attention on how individuals interact one with each other (micro-level) under conditions already established in the past but that cause the social phenomena (macro-level). The most required methods used to unify the micro and macro level are differential-equation models and laboratory experiments. From these models, we can derive four conclusions on the relationship between the individual actions and social phenomena:

1. The size of the cause and effect are not necessarily proportional.

2. The structure of social interaction has its importance for the social phenomena.

3. The effect that an action (micro level) has on the social (macro) may be related to how the structure to which the actor is due.

4. Aggregate patterns do not give enough explanation about the micro-level processes in charge of realizing the macro level. These four conclusions conclude that the relations between the individual and the social are difficult, thus as it was mentioned before, the outcomes are complex to predict without analytical tools.

Chapter 5. The Schelling Model

The Agent-based model always trying to explain complex social phenomena. To put an example: we have the ants, ants are simple, they do not have a brain but can construct very complex structures. This means that sometimes you do not need complexity at the individual level to have complexity at the social level. The segregation model explains this very well. Thomas C. Schelling is the author of micromotives and macrobehaviour. The Schelling´s game will model the segregation process. This game has several assumptions. The first one is that there is an initial state- residential segregation- with an unidimensional town, where there are to pieces in a line situated from left to right. The second assumption, there is a neighbourhood that consist of the, household itself, the four households to the left and the four households to the right. Now we know that segregation is a dynamic process, because it changes over time. So the third assumption will be based of moving households. Where the game proceeds in rounds. So one by one, from left to right, each household does the following:

1. If more than half of the neighbours, the majority have the opposite colour, the household becomes unhappy.

2. If a household is unhappy, it moves to nearest available location where it would not be unhappy. Finally, there are two outcomes, the end of the game would be no household is unhappy or there are no suitable places for unhappy households to move in. So we have a very simple situation, simple decision-rules, no intrinsic preferences for segregation, all they want to do is to avoid the minority. But the final outcome is ending up in a very extreme situation with high levels of segregation. But which kind of model is the Schelling´s model? It is not desired and belief mediator but opportunity mediator. And how does this Schelling´s model differs from the ¨common sense theory¨? The model itself shows high levels of segregation can result from rather tolerant individual preferences. High levels of segregation can result from both tolerant and intolerant preferences. Plus Schelling makes the micro-macro link explicit. However, in the common sense theory says nothing about how segregation happens. Does not mention anything about this micro-macro interrelation. Finally the Schelling model allows for derivation of new implications.

In conclusion the Schelling´s model is unrealistic and simplifying assumptions on:

1. one-dimensional town

2. Order of moving

3. two groups of similar size

4. homogeneity assumption with respect to neighbourhood sixe and preferences of neighbourhood composition

5. No scarcity on the housing market

6. No legal or economic constraints.

So that is why we do need a better model. A two-dimensional variant, it is a game where we have two cities, more actors, both groups do not want to be in the minority situation. Assuming the two is a linear line, we still ended up with high levels of segregation.

What else do we learn from this Schelling´s model? It is a complex social phenomena that is brought about by simple individual behaviour. The collective outcomes are not a on-to-one reflection of individual preferences. In order to predict a final outcome, we make use of a theoretical tool such as a computer simulation, which will allow us to study the micro-macro link. The aim of this simulation would be to derive implications from complex models when formal analysis is problematic. It is perfectly suited for social science since it allows studying the macro-level behaviour of systems of many interacting actors. Some famous actors using this simulation are Schelling (1971), Boudon (1976), Club of Rome (1979), Axeldrod with its model of evolution.

Chapter 6. Quantitative Research, Agent-Based modelling and Theories of the Social

There is a lot of macro phenomena which are hard to explain and predict. Examples of macro phenomena can be mass collective actions (demonstrations, riots, revolution and so on). Another examples of puzzling macro phenomena are the Diffusion of beliefs, innovations; Fads, hypes; Economic ¨bubbles and Emergence of social structure.

Now what are the problems that came out when explaining social change? We do know that social changes can happen suddenly, in an unpredictable way. The explanation by simple aggregation of individual preferences, change of norms, is not always enough. So that is why social changes are likely to be unintended consequences of individual behaviour. The coleman boat will help us to explain the complexity of transformational mechanism. We will be able to understand how individual´s actions get together and lead to certain collective outcomes. This can be considered to be the most risky task for sociologists. And that is the reason why analytical tools are usually needed.

The model for this type of phenomena must refer to Mark Granovetter. He is one of the ¨headmasters¨ of social network analysis. In concrete, his most known theories are based on ¨the strength of weal ties¨ and ¨the problem of embeddedness¨. Now we will focus on Granovetter´s threshold model of collective behaviour. This model is motivated to explain riots, demonstrations, etc. The assumptions count with ¨N¨ number of actors, each of these actors have a binary choice. The cost and benefits depend on how many other individuals join or not. However the costs and benefits differ between actors. That is why we say each actor has his threshold. The process takes place in rounds and finally actors are goal-directed. They do always want to maximize utility.

What is the threshold of the actor? It is the proportion of the population who have to join before another actor does the same. If an actor always joins, the percentage will be 0, if more than 50% or more other actors did join, then the actor will join as well and his threshold will be 50%. If the threshold is 100%, the actor never joins.

The structure of the threshold model starts with an initial behaviour, distribution of preferences, after actors observe all others, they will measure cost and benefits of alternatives, as it is mentioned before, the individual´s threshold. This individual has a goal-direct action. When all individual actions pursue a same goal, we will end up with a collective behaviour. As analytical conclusions to this threshold model, we can say that collective behaviour can change in a sudden, for a given distribution of thresholds. At the same time, this collective behaviour can be extremely different between the different distributions, even if the mean of the distribution does not change (see book for example). The methodological conclusions to the model says that small changes at the micro-level can lead to bigger changes at the macro-level. This macro level is complex and is brought about by simple assumptions at the micro-level. That is why we need analytical tools to understand the complete behaviour of the model.