The main aim of this thesis is to investigate the models social scientists construct about the world as they see it. Specifically I am asking "Did Paul Stirling in his analysis of social change in a Turkish village over a 42 year period 'get it right'"? "What does it 'mean' to get a 'social model right'"? In trying to address these questions I will be concerned with the causes and effects of labour migration22 in Sakaltutan and the 'intuitively obvious'? but highly abstract concepts of causality, modeling & 'causal change in societies'.
Stirling's data is unusual in its quantity, quality and the time period over which it runs. For these reasons it is suitable for a range of analyses rarely undertaken. Stirling has provided not only unusual data but has also brought to bear a consistent and cumulative analysis of it. He has formulated a set of 'models' of social processes and, although it is beyond the scope of this dissertation to evaluate them all, I will attempt to evaluate the descriptive and explanatory power of one of these models.
As this was a preliminary investigative study, far more data was analysed and supporting materials collected than is necessary for the arguments outlined below. For these reasons a lot of the Supplementary data and analysis are held in appendices on the web site version. of this dissertation.
In Chapter One I will present a review of the relevant ethnographic materials. I will outline the development of Stirling's ideas on causality and examine the 'models' he built up to help explain his findings. I will introduce his now infamous 'metonymy'? "The Wiring Diagram" and look specifically at his models of population change & migration. Then in Chapter Two I will discuss the use of standard computer based tools i.e. databases, 3D graphical & statistics packages used in quantitative and qualitative analysis as well as more esoteric, bespoke, mathematical modeling techniques such as simulation. I will provide some theory as to why these methods might be useful. In Chapter Three I will analyse some of the data in Stirling's Turkish Village database and, producing population demographics, and other primary indexes of change within the village, present a simplified proto simulation model of migration. It is the eventual aim of this project to investigate Stirling's models further by the use of more advanced simulation techniques. In Chapter Four I will summarise the results, namely that social science models can indeed be "tested' if the ethnography is good enough. When well done, the art of modeling can penetrate into the mysts of the social world and a developing 'systems' approach, although initially demanding of the practictioner, can offer new insights, confirmations or rebutals of anthropological work. I will conclude by discussing the web site version of this dissertation and its implications for collaborative research.
The Ethnographic Context
Field Work & General Overview.
When Paul Stirling died in June 1998 he left behind him an extensive and impressive ethnography of his life's work on Turkish villages including his and other researchers field notes, a book, published articles, unpublished drafts, book reviews, newspapers and periodical pieces;photographs, four films, a website  and a database. The following ethnographic review will use these resources, most of which are also freely available to other researchers via the internet, to trace the diachronic development of two particular lines of his thought, those concerning labour migration and causal processes.
On the web site version of this dissertation there are hypertext links to most quoted documents to allow for more detailed comparison.
In 1939 Stirling read and was inspired by Ruth Benedict's 'Patterns of Culture' after an Oxford year of undergraduate Latin & Greek. He saw active service as a Royal Navy Officer during the war and returned to Oxford in its aftermath to complete a first degree in Ancient History and Philosophy. Dissatisfied with the questions raised in these disciplines he decided to become a social anthropologist. He was fascinated by the idea of comparative studies of societies, especially the concepts of social systems & social structures. He left to do field work in Turkey, funded by the British Treasury, in 1949 with this advice from his supervisors:
Meyer Fortes said 'Everything in the village concerns you'. Evans-Pritchard said 'Find what they are most interested in and study that.'(Stirling 1992b)
He was looking for 'social structure'; groups, networks and hierarchy and has reflected since that he had not realised how much a 'mildly informed but detailed ethnography of (the religion -Ed.) might contribute'(Stirling 1992b). Nor did he expect to be able to understand the way the villagers thought and felt with only an outsiders command of language (Stirling 1992b). He says that he was 'aware but suspicious of 'culture' and of grand evolutionary theories' and perceived himself to be 'somewhat arrogantly philosophically cynical' (Stirling 1992b). He also had doubts about 'structure' & 'social system' and about the way historians and Islamists wrote about history and Islam.
In April 1949 he arrived together with his wife Margaret in Turkey. Their intention was to study villagers simply because he believed that this was where 'our' greatest ignorances lay (Stirling, 1965). He had decided to study one village thoroughly and another in less detail.
His choices of villages were more or less accidental. When the staff of an American school and clinic situated near Kayseri offered them a base, he accepted. The village he chose to study with the most detail was called Sakaltutan which means 'beard grasper'. The villagers tell the story that: 'a man caught his beard in a narrow pass on the road across the bowl from the village'. He later discovered that Sakaltutan was the name of a tribe / lineage from Asia and that beard grasping is quite an aggressive act. However, the village is thought to be less than 200 years old and it is likely that is was settled after having been used as a pastural summer camp by people from villages near Kayseri.
He was in Sakaltutan from November 1949 to August 1950. Margaret joined him in the March of 1950. Sakaltutan had 'an excellent water supply, more or less regular lorries into Kayseri, and eleven other villages within an hour and a half's walk' (Stirling 1965).
During their time they managed to collect a simple house by house census & make notes on genealogies. Both he and Margaret produced comprehensive field notes and took B&W Photographs
In 1951 they returned to the area and from August to November lived in a second village called Elbasi some five hours walk east of Sakaltutan. Elbasi was studied because it provided a contrast to Sakaltutan, being richer, more agriculturally dependent, more sophisticated, less isolated and in the center of the main district of Nahiye. He reflected later that it was the similarities between the villages that in fact that were the most striking.
At the end of 1951 he returned to the UK to take up a post at the L.S.E.but returned to Elbasi alone for two months during the summer in 1952. His fieldwork monograph 'Turkish Village' was finally published in 1965.
In 1971 he returned for the first time to Sakaltutan for two months but because
of the military government, was only able to do two weeks work in the villages with
the help of an official 'guide'. He only had three days in each. He managed to collect
information on every person in his 1951 census in Sakaltutan by questioning a constantly
changing room full of people who came to greet him. In Elbasi he was only able to
check up on every person in every third household from his 1951 census. His records
reflect his astonishment at the changes in village life he encountered.
From 1974 onwards he was able to visit Turkey every year apart from 1977. From 1983 to 1986 he was visiting Professor at the Middle East Technical University in Ankara. He held official research permission from July - October 1985 at which time it was suddenly rescinded without any explanation, however from March 1985 he was able to collect data on all individuals and households in Sakaltutan and about half from Elbasi by driving 14 miles by car from Talas every day armed with specially printed interview cards. In 1989 he returned to Turkey for six weeks between June and July, and from April to June 1990 to check on and expand his data. His last publication directly concerning the villages was in 1994.
In this section I will try to outline some of Stirling's ideas on social causality and illustrate the kinds of data he was collecting.
Stirling's early work reflects a strong interest in structural analysis. In his first articles 'Social Rank in a Turkish Village' (Stirling, 1953) and 'Land, Marriage, and the Law in Turkish villages '(Stirling1957) which are based on his 1949-50 field trips, he gives basic statistics on population and household numbers & estimates of landholdings in the villages. He describes village ownership of land, marriage and inheritance(see Appendix B), details of staple crops grown44and times of harvest5 approximate yields and available agricultural technologies, transport and communications.
In his analysis of 'rank1' (Stirling, 1953) he comes to the 'general' conclusions that within Sakaltutan there were:
1. No formal hierarchical schemas.
2. No hereditary stratification.
3. No marked etiquette of privilege.
He states that the most important ranking 'scale' is wealth, in this case land and animals.
Stirling's paper published in 1963 The domestic cycle and the distribution of power in Turkish Villages is a seminal paper. It offers valuable structural generalisations of the village and is fundamental in that he outlines, giving historical evidence, two possible models of land availability. He identifies 2 main characteristics of the domestic cycle and 2 main characteristics of the distribution of power within the village(Appendix B2).
A full joint 3 generation household type is posited as being the 'ideal' (Stirling, 1963). The average age expectancy he quotes as around sixty for men with an average gap of about 25 years between father and son. Therefore, an average household head would have to spend around 15 years after his father died before his grandson was born leaving him ten years as household head before his own death. So, in an average village at any one time less than half the households would be full joint ones. (Stirling, 1963). In reality he reports that only 15 out of 105 house holds fulfilled this ideal. This contained 22% of the population. Obviously achieving a full joint household was difficult. Prior to 1950 the infant mortality rate had been high, approximately 30% (Stirling, 1998) as well as the infertility rate. Miscarriages were common. Couples could conceive a run of daughters. Adult men died prematurely leaving unmarried children. This had been accentuated by events such as conscription into the Ottoman army. The wars in turkey between 1911 and 1922 had a marked effect and he states that in the past the death rates had been even higher.
General historical notes of interest are provided in the notes section23.
As a result of the second world war many village men had seen active service and the village population had again suffered from their absence. This is borne out by the demographics. Their return however brought knowledge of new medicines. In the early 1950's the U.S. was pouring aid into Turkey in the form of cash, medicines and infrastructure i.e. roads for U.S. military purposes. These new roads ran close to the remote villages whose menfolk, often with picks and shovels, dug feeder roads to them. The newly elected government was also trying to implement elementary health care policies. In this way goods such as medicines were making their way into the villages and being accepted for the first time. Margaret and Paul Stirling were present in the village when the very first antibiotics arrived. They boiled up the penicillin in bowls on top of billy cans before administering it to the very first patients for tuberculosis, bad cuts and fevers etc.
In a prophetic statement Stirling says:
It is apparent with hindsight that Stirling's analysis was essentially correct. He had managed to analysis the social dynamics to the point of being able to make simple conditional predictions after having only been there at this stage for less than a year.
Marriage patterns are an essential component of the village lifecycle and will be used in more complex analysis in the future.There seems however to have been little in the way of stratified marriage arrangements. I have summarised the data in Appendix B5.
1950 seems to have been a water shed in the history of Sakaltutan and of Turkey Generally11 (Stirling 1974). It is probable that the village had being undergoing constant changes over the last 200 years but its environmental 'constraints' were about to be changed in ways previously impossible.
After the general election in 1950 the new democratic government showed conspicuous interest in maintaining the villagers vote by providing services such as roads and water supplies. Public buildings and schooling were also taken over by the state. The government provided the village also with a little paid employment. The state office of soil products which controlled grain prices & agricultural credit was already taken for granted at this stage and the two areas of 'need' lay in improvements in technological efficiency and health care - cures for current illness. Around 1950 the government was trying to provide heath care visitors for the villages.
1n 1950 Stirling had classified village households into 4 groups: (Stirling, 1965 p69):
49 depended entirely on their land, or used other sources only to a minor degree in
16 derived the main part of their income from work outside the village;
22 normally used both these sources;
13 derived their income largely or wholly from work inside the village.
It is clear that by the early 60's this is starting to change dramatically45 (see Appendix C).
By 1965 cash was becoming the preferred means of meeting obligations within the village. Permanent migration was still however quite rare and by all accounts most people preferred to stay in the village with people they knew.
From a passage on gender differences: 12
we can see that at this stage only men could ever migrate to work. The women could only generate income if it was within the village. Since the 1960's Sakaltutan had been part of the home working sector of the Kayseri hand made carpet industry with practically all village girls weaving during the winter for very little pay. What pay they received would probably go towards their 'Trousseau' or bride gift or to the head of the house hold.
The first permanent migrants, in accordance with national trends, moved out with their whole households and settled in town around 1957 (Stirling, 1988). He argues that the village was being transformed into a small agro-business. By the early 70's agricultural labour had dropped dramatically with the introduction of tractors and fertilizers and there was more surplus to sell.
By 1988 he states that a patrilineal reckoning of household growth gives a figure of 3 households for every one in 1950 and estimates that about half of these descended house holds are now in town. (Stirling, 1988)
He gives the average house hold size for 1950 as being 6.0. 6.4 in 1971 and 5.74 in 1986. In the towns in 1986 this was 4.25, up from 1971 and climbing.
In Labour Migration in Turkey: Thirty Five Years of Changes, 1991
He continues his analysis of the census data and includes a table for Sakaltutan showing the increase of population and the rate of migrant households outflows from 1950 - 1986. (Appendix C)
Stirling reports that by 1993 the outflow of households from Sakaltutan had increased and only about 110 were left.
By working with his new database - still not completed at this point, he was able to compute other useful indexes such as income breakdown estimates for households from 1950 - 1986. (Table 2, Appendix C)
Stirling & Áncirlioglu, (1994) are able to give a diachronic breakdown of villagers with migration experiences over the periods of his census(see Appendix C). This shows a rise from 48.6% in 1950 to 86.9% in 1986 for males over 14 years of age and a rise of 24.7% in 1950 to 50.7% in 1986 for females over 14 years of age. Marriage to first cousins has been and still is common in Turkey but Stirling's field work from 1971 to 1986 shows a '24 per cent in 1951 to 1960 to 29 per cent in 1980' (Stirling, Áncirlioglu, 1994) and it was still increasing in 198628-31. They suggest that this is more to do in risk evaluation and changes in social cognition than any form of Straussian alliance theory26.
Land in the village had always been of principal importance. Not only land but Oxen to work the land. After having argued that legal changes at the national level regarding land reform are only indirectly relevant to the village (Stirling, 1957), and therefore that he is justified in modeling the recent village history independently of national reforms, he outlines two models of land availability. The conditions for ploughing are similar in each.35
Type 1. Surplus of land - More land is available for cultivation than the village population with their technical resources can use. Hence anyone who has the necessary manpower, draught animals and seed, can plough land freely. What emerges is a self regulating social system, constrained by its natural environment, with little in the way of hereditary social stratification in which it is the hard work, vitality and 'luck' of each household as to how they fair from generation to generation.
Type 2. Scarcity of land - All cultivable land is under cultivation, except for essential meadow land for pasture. In this situation prestige and wealth can no longer be obtained by simply being able to 'plough more land'. Other avenues have to be sort in which a household can be 'successful'. It is no longer the begetting of sons that necessarily leads to status and wealth and reluctant migrations to more fertile pastures soon follow.
I summarise the relevant details for Type 1 Surplus of land in Appendix B3. (Stirling qualifies his statements here carefully6).
A summary of the details of model Type 2, Scarcity of land is in Appendix B4.
Stirling claims that by 1950 land was already scarce and the possibilities for increasing 'wealth' limited.(Stirling, 1953). He suggests that the possibilities for earning cash outside the village were increasing and by having sons as migrant workers the household head could still become wealthy. This is important as status was not hereditary (Stirling, 1953). Each man had to rise from the 'poverty' of his share of inherited land from his father to a position of wealth 'on the backs of his own sons' (Stirling, 1953).
Migrant workers however were not dependent on their fathers for their earnings and could leave home if totally dissatisfied. So migrant workers could both help increase the status of the household but at the same time undermine the authority of the father.
Stirling has identified two influences on the dynamic of the village as being the need of the sons to find their own wealth and status, and the availability of resources that allow them to do it.
He was fully aware of the exceptions to any general statement he could make but nevertheless in doing so he came up with some 'rules' & 'constraints' implicit in these two situations. The analysis of them lead to his first explicitly stated 'cyclical feedback model':
From these models Stirling forecasts in 1963 the following:
To the extent that these two model might hold, he suggests that the important transitional period between them was in the late 1940's when there was a 'land rush'. Villagers vied with each other to claim 'common land' by ploughing it up. Stories abound of competing lineage and household members ploughing strips alongside each other in order to render claims useless. (Stirling, 1965).
These two forces collide when land is in shortage. If migration is an option and land is scarce then here is a motivation for a son to take that option - even if the land can still carry the population.
If we try to ask then what really was going on in these villages at that time37, Stirling's observations that the state's new system of land and marriage registration4 was not taken seriously by the villages (Stirling, 1957) renders the official statistics, at best unreliable, and at worst useless. In order to go further we need the kind of data that only an immersed longitudinal study can provide.
Stirling is obviously already used to 'manipulating' his mental models when he speculates about the consequences of a later or earlier division of the household9 and about the relationship between inherited power and this domestic cycle10. From these 'simulations' he can see how the system oscillates and he makes an important remark concerning social stability.
His ability to analysis such complicated systems were not without reward. In his 1963 article he projects his analysis towards the future and anticipates much of the development of the next 20 years & he felt that he understood the situation well enough to make the following theoretical observation.
There is evidence that Stirling's thinking has been increasingly occupied with notions of causality in his 1970 article for The New Atlantis:
and in an important passage he says (my bold print):
He is trapped between 'knowing' on the one hand that a good analysis has a degree of prediction in it, and having the understanding that the social world is so complex that any attempt to represent it in terms of relationships and variables seems absurd. Following up to this he then argues:
and in a hark back to sentiments expressed in his 1953 paper, he asks:
His return to the field in 197118 and his dissatisfaction with the explanatory theories he was familiar with seem to have been the motivating spur for his next and arguably most important workCause, Knowledge and Change: Turkish Village Revisited (1974)
This article appeared four years later in a book dedicated to the anthropologist Lucy Mair, "Choice and Change."J.Davis (Ed.) (1974).
It is very obvious19 that he felt a real need to address these problems of 'representation'. He goes on to ask:
By problematising the situation in this way he was able to attempt an answer. He takes as his baseline Lucy Mair's 1969 article'How small-scale societies change' and extends her definitions.
He proposes four types of change.
1. Social Relationships - role relationships between people living in the village and in nearby villages, and also those with people outside the rural area--with officials, employers, fellow workers, bank managers.
2. Knowledge and Beliefs21 - this he concedes is a difficult category to define20 and make the crucial point:
What people believe is....extremely difficult to state accurately. Yet to describe social change, we must make the attempt, since beliefs certainly change; and change more than people are normally willing or able to make explicit.(Stirling, 1974)
Of this category Stirling says:
If peoples knowledge expands greatly, then there are likely to be changes in the notions of what to aim for, what is desirable, what is right and good, what is tolerable, what is acceptable and what is unacceptable.(Stirling, 1974)
4. General Characteristics
In this 'everything else' bag he mentions the effect of the introduction of money & market principles on traditional reciprocities and the move towards bartering & profiteering. He notes that if the possibilities of acquiring knowledge, skill & wealth are greater then changes in the village hierarchy will accelerate and that external intervention in the form of state and commercial institutions not only effect the village but blur the distinctions between village and the 'outside'.
He then tries to conceptualise 'Cause':
and in the next paragraph he states explicitly for the first time his theoretical aim:
At the end of many years of chewing the cud he has started to crystalise ideas that have been bothering him for years. This paragraph is the key to understanding his motivation for producing his infamous 'wiring diagram'. It is a transitional paragraph which in thinking terms leads from the window of the poet to the door of the mathematician. This diagram will occupy the rest of this dissertation. The boxes and arrows of specific interest are marked in RED.
THE 'WIRING DIAGRAM'
In Stirling's words:
The main diagram summarises some of the ideas that occurred to me while I was in the field, and afterwards while going over my notes. I can weight neither factors nor causal links; the factors are not mutually exclusive, nor are the boundaries around them always clear. These 'factors' indeed are not even things of the same general order of existence. Equally the causal connections for which the lines and arrows stand are not of the same kind, and do not reflect similar 'efficacy' or similar processes. Even the double boundaries which I use to emphasise certain factors as crucial are selected on my own judgement.....Any two-dimensional diagram must simplify reality. Even so, the point of the diagram is to illustrate complexity, and these misgivings I make explicit in order to add weight to this emphasis. In what follows I shall not discuss all thirty-one boxes, for some hardly need comment; nor will they always be discussed in serial order, but rather in the order which best suits the development of my argument.(Stirling, 1974)
A full explanation of the various boxes is available on the website and in Appendix B6.
In the section 'Some Causal Chains' he suggests that the connections be looked upon as 'hypotheses' and illustrates how causal chains can be extracted from the diagram.
He extracts three casual chains40 all of which are relevant to migration and population growth.(def. Permanent & Pendular migrants)39.They are reproduced below:
Causal Chain 141
Falling death rate
Less village resources per head
Better nutrition and better medicine
Falling death rate
It is this chain that I will refer back to in the later analysis.
Causal Chain 242
Sporadic pendular migration
Higher minimum standards of living
Aspirations for higher incomes
Regular migration for all young men
This model simply shows that, given other factors, once sporadic pendular migration has begun, it causes a rise in village expectations, which in turn causes migration to become a normal and regular part of village life.(Stirling, 1974)
Causal Chain 343
Increased pendular migration
Increased permanent migration
These simple chains are examples of feedback loops (Forrester, 1968) in which a limited number of variables effect an outcome in a limiting, accelerating or stablising fashion. They are the basics of any form of systems dynamics analysis. Stirling, apparently with little knowledge of these fields has followed his 'causal instincts' and come up with pretty much the same basics.
For the first time he has drawn together his thoughts and concerns from the previous 25 years into a theoretical paper but he is still problematised by the result. His levels of insight & analysis at this point are impressive and so is the manner in which he choose to encapsulate this 'understanding' 'The wiring diagram' is a simple 'metonymic device'. A 2 dimensional diagram showing relationships between conceptual entities abstracted from his own constantly refined, reassembled, reworked ideas and experiences of Turkey and the villages. The fact that it appears so simple is both its mastery and its mystery. To the analytically shallow and less engaged mind its attempt to convey complexity through simplicity is laughable, but to the initiate it provides a pathway through a very tangled forest of understanding to the beginnings of a realisation. This sort of causal thinking, now unfashionable in the social sciences is I feel at the very heart of the scientific endeavour. Not as a simplistic 'reductionist' hard assertion of reality, but rather as providing a clear spotlight through a classificatory jungle to the things that matter. It provides insight by illuminating 'patterns in the mist'. It is all too popular currently to believe that the mist is all that is. This belief of course, as Alan Sokal points out, has consequences as well!
In summary then Stirling arrived in Sakaltutan at what is now regarded as a water shed in Turkish history - 1950. The village had already become partially dependent of migrant labour but few households had moved permanently. Population was growing because of a fall in the infant mortality rate whilst the birth and death rates were pretty much the same. This increase in population was putting pressure on village resources with the result that households started to migrate out as well as individuals. Stirling was well aware of the complex and multidimensional aspects of cultural change and his 'wiring diagram' is a way of trying to represent what he thought the main factors were that were changing life in the village. The question is does it have more power than being just one mans subjective map full of his own purely realtivistic notions? I will now proceed to analyses a small part of this 'metonymy'.