|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
CULTURAL MODELS AMONG TRANSNATIONAL MEXICAN MIGRANTS
MARK C. HOUSE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Mark C. House
To Papa, Mom and Lisa
Assistance comes in many forms. Some aid requires more effort or time or resources but
all efforts were crucial and I am grateful to every person who was willing to help.
Over the past few years I have had the opportunity to know a number of fellow graduate
students and a few undergraduate students at the University of Florida who have been very open
minded and willing to help. They provide a sounding board for my thoughts and a critical eye to
help me better understand the ideas I encountered. Perhaps the most influential of these students
have been the group known as SOR (Students of Russ). This group includes a number of Russ'
students who have been especially helpful, including; Amber Wutich, Rosalyn Negron, Stacey
Giroux, David Kennedy, and Lance Gravlee. Theses individuals were the support group that I
needed to complete graduate school and to better understand the ideas and methods we were
There are only a few times in a person's life that you have the opportunity to work with
people that are as accomplished as the professors at the University of Florida. While some stick
out, every class I took was a challenge. They are exceptional scholars and patient teachers. Dr.
Tony Oliver-Smith, Dr. Michael Heckenberger, and Dr. John Moore have been especially
inspirational. Dr. Stan Smith has also been very helpful and patient in allowing me to work with
him on the publication of two articles. I also have the deepest respect for my committee members
- Dr. Chris McCarty, Dr. Anita Spring, and Dr. Chris Janiszewski. I have admired their
professionalism, their teaching ability, and their insight. I have the greatest respect for my chair,
Dr. H. Russell Bernard. He has provided the guidance and patience for me to be successful at
Levi Rios Cardenas, Ananda Rios Cardenas, Abraham Rios Cardenas, Maria Rubio Inda,
Antonio Rubio Inda, and Beatriz Cardenas were kind enough to share their home and their lives
with me. Mildred Denisse Arce Zuno is a wonderful Spanish teacher and good friend. There are
also a number of people who assisted me with my research. They were patient and hard working
in their efforts. This group includes Jairo Rodriquez Rubio, Arturo Ruiz Licon, liliana Elizabeth
Vega Sandoval and Jessica Elizabeth Solano Lopez. The faculty at the San Isidro Mazatepec
Prepatoria were also very helpful, especially Senior Jorge Aceves Mercado. In the U.S. there
were also a number of helpful individuals who introduced me into their community and assisted
with the data collection. These include Stephen Mardock, Kelsey M. Leak, Valerai Luna, Crystal
Hernandez, Nicole Roeder and Latisha Larson. There are a number of people who pointed me in
the right direction or helped me gain access into an institution. Without them this proj ect would
have taken much longer than it did. For this assistance I am grateful to Paul Chaffin and Brenda
Chatfield at the Kansas Northwest Area Vocational-Technical College; Lowell Coon, Linda
Davis-Stephens, and especially Larry Koon at Colby Community College; Harvey Swager, the
principal at Goodland High School; and Mabel Burciaga, Margie Hernandez and Duane Arntt at
Burlington High School.
My work in graduate school is only possible with the support of my family. My parents
have always encouraged me to pursue my education. My grandfather supported me when others
had doubts. However it is my wife, Lisa, who has contributed the most. Not only has she asked
the well-reasoned questions only an economist can ask, but she also provided a release for my
frustration and a reminder that there is light at the end of the tunnel. She allowed me the time to
take classes, conduct my fieldwork and finally to write this paper.
There are two people that, not only do I need to thank, but to whom I also need to
apologize. I apologize because they made a sacrifice they didn't know they were making. In
completing this proj ect I regret that I gave up precious time with my daughters, Kassidy and
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. ...............4.....
LIST OF TABLES ............ ...... .__. ...............10...
LIST OF FIGURES .............. ...............12....
AB S TRAC T ............._. .......... ..............._ 13...
1 INTRODUCTION .............. ...............15....
Why Study Migration? ............ ...............16.....
Why Use Cultural Domains? ............. ...............18.....
Manuscript Organization .............. ...............20....
2 MEXICAN MIGRATION ................. ...............22...............
Mexican -U.S. Migration and Policy .............. ...............23....
Early Policies ........_................. ..........._..........2
The Bracero Program .............. ...............24....
Post War Immigration .............. ...............25....
IRCA ................ ... .... .._._ ...............26......
Theories of Why People Migrate ................. ...............28....... ....
Early Equilibrium Theory............... ...............29.
Neoclassical Economic Theory .............. ...............30....
Dual Labor Market Theory ................. ...............31....... .....
New Economics of Migration ........._.. _........... ...............32..
Dependency Theory............... ...............32.
World Systems Theory ................. ...............34.............
Articulation Theory .............. ... ............ ...... ..........3
Summary of Factors in the Creation of Migration Streams .............. .....................3
Why Does Migration Continue? ............. ...............36.....
Other Factors in Migration .............. ...............38....
Racism .............. ........ ... ..... .........3
Temporary Versus Permanent Migrants ..................._.__ ......... ............3
Le gality ................ ...............40..............
Remittances ...................... ...............41
Labor Organization of Migrants ................. ...............42....... ....
Post Migration Changes ................. ...............43.............
Assimilation and Acculturation ............ ..... ..__ ...............44...
T ransnationali sm .............. ...............47....
Hypotheses ........._... ...... .__ ...............54...
3 CULTURAL DOMAINS .............. ...............55....
Cognitive Anthropology .............. ...............55....
Beginnings ............... ...............56....
The New Ethnography ..................... ...............57.
Semantic Networks and Ethnobotany ................. ...............60................
Prototypes .............. ...............61....
Schema Theory .............. ...............63....
Cultural Model s .............. ...............65....
Cultural Consensus ................. ...............67...
Cultural Models and Behavior............... ...............68
A Critique of Cognitive Anthropology ................. ......... ...............70.....
Hypotheses ................. ...............72.................
4 METHODOLOGY .............. ...............74....
Research Design .............. ...............74....
Sampling .............. ...............76....
Locations .............. ... ........ .............7
San Isidro Mazatepec, Mexico ...................... ...............77
Northwest Kansas and Northeast Colorado, USA ................. .............................78
Sampling Strategy .............. ...............8 1....
Research Methods ................. ...............83.................
Participant Observation .............. ...............83....
Mexi co ................. ...............8.. 3..............
Americans ................. ...............84.................
Mi grants ...................... ...............8 4
Assistants and Training .............. ...............85....
Freelists and Pilesorts ................ ...............86........... ....
Interviews .............. .... ...............88..
Demographic Information .............. ...............88....
Cohen' s Transnational Score ........._..... ...._... ...............88....
ARSMA- II ............. ........... ...............90.....
5 DOMAIN ELEMENT S .............. ...............95....
Dataset Summary............... ...............96
Domain and Population Names .............. ...............96....
Cleaning the Freelists .............. ...............96....
Qualitative Analysis............... ...............97
Domain Composition ........._.._.. ....._.. ...............100.....
Domain Overlap Analysis I ........._.._.. ...._... ...............100...
Domain Overlap Analysis II............... ...............103..
Summary of Domain Overlap Analysis .............. ...............105....
6 DOMAIN STRUCTURE............... ...............12
Domain Breadth ................. ...............126......_ ......
Raw Freelist Length .............. ...............126....
Average Freelist Length .............. ...............127....
Variance of Domains ................. ... ............ ...............128.....
Testing for Differences in Freelist Length .............. ...............128....
Discussion of Freelist Length ..............._ ...............129..._... ....
Domain Structure ........._.._... ....... ...............129....
Cultural Consensus Analysis ........._... ...... .___ ...............130....
Alternative Domain Structures ........._._._. ....___ ...............131...
Subcultures within the Domains ........._._._ ....__. ...............133...
Discussion of Domain Structure ........._._._ ....__. ...............134...
Factors Affecting Structure ........._... ...... ..... ...............136...
Domain Knowl ed ge ........._... ...... ..... ...............136....
Demographic Variables ........._... ...... .___ ...............137....
Transnationalism .............. ...............137....
Acculturation .............. ...............138....
M ethod ............ _...... ._ ...............139...
Results ............... ..... ......__...... ...___ ... ..........14
Discussion of Factors Affecting Structure .............. ...............141....
7 DI SCU SSION ........._.._.. .. ............... 15....._. 5.....
Summary of the Analysis of Domain Elements ........._.. ............ ......__. ........15
Summary of the Analysis of Domain Structure ................. ....._.._............ .......15
Summary of the Analysis of Domain Knowledge ................. ...._.._ .............. .....15
DEMOGRAPHIC QUESTIONAIRRE AND COHEN' S TRANSNATIONAL SCALE...........159
SPANISH ARSMA II SCALE .............. .....................160
LIST OF REFERENCES ............ ..... .__ ...............162..
BIOGRAPHICAL SKETCH ............ ..... .__ ...............183...
LIST OF TABLES
Table 4-1 Number of responses collected by type and population ................. .......................91
Table 4-2 Number of items selected from freelists ................. ...............92........... .
Table 4-3 Participant demographic c information ................ ...............92...............
Table 4-4 Cohen individual transnational scores ........._..._. ....._... ....._._...........9
Table 4-5 ARSMA-II scores ................. ...............94......... ....
Table 5-1 Freelist response totals .............. ...............106....
Table 5-2 Open pilesort response totals ........._._.._......_.. ...............107..
Table 5-3 Constrained pilesort response totals .............. ...............107....
Table 5-4 Things mentioned by more than 20% of the respondents .............. .....................0
Table 5-5 Foods mentioned by more than 20% of the respondents............... ..............10
Table 5-6 Pastimes mentioned by more than 20% of the respondents ........._.._.. ......_.._.......1 10
Table 5-7 Unique foods ........._.._.. ...._... ...............113...
Table 5-8 Unique pastimes ........._.._.. ...._... ...............117...
Table 5-9 Unique things............... ...............120
Table 5-10 Number of items in the reduced datasets ........._._.._......_.. ...............123
Table 5-1 1 Percentage overlap of domains for complete and reduced datasets .......................... 125
Table 6-1 Total freelist length............... ...............142
Table 6-2 Average freelist length .............. ...............143....
Table 6-3 Standard deviation of freelists ............ ........... ...............143..
Table 6-4 Significant of T-test for differences in mean freelist lengths of foods ................... .....143
Table 6-5 Significant of T-test for differences in mean freelist lengths of pastimes .................144
Table 6-6 Significant of T-test for differences in mean freeli st lengths of things .....................144
Table 6-7 ANOVA results food ................. ...............144..............
Table 6-8 ANOVA results pastimes .............. ...............145....
Table 6-9 ANOVA results things .............. ...............145....
Table 6-10 Consensus analysis eigenvalues for each domain and sample .............. .................146
Table 6-11 Comparison of eigenvalue ratios for people who have children .............. ..... ........._.147
Table 6-12 Comparison of eigenvalue ratios for people who have 9 years or more of
education ................. ...............147................
Table 6-13 Comparison of eigenvalue ratios for females ................. ...............148............
Table 6-14 Comparison of eigenvalue ratios for people who are married .............. ..................148
Table 6-15 Comparison of eigenvalues for people 25 years old and younger............................. 149
Table 6-16 Summary statistics for the knowledge scores in each domain .............. .................152
Table 6-17 Summary statistics for the standard and modified transnational score .....................152
Table 6-18 Summary statistics for MOS, AOS and ARSMA-II scores ................... ...............152
Table 6-19 Model summary pastimes .............. ...............153....
Table 6-20 Coefficients pastimes .............. ...............153....
Table 6-21 Model summary things............... ...............153
Table 6-22 Coefficients things .............. ...............153....
Table 6-23 Model Summary foods ................. ...............153..............
Table 6-24 Coefficients foods .............. ...............154....
LIST OF FIGURES
Figure 4-1 Overview of research design ........._._. ..... .__ ...............91..
Figure 4-2 Frequency of freelist items for migrants ......................_. ...._.... ...............92
Figure 5-1 Overlap of domains for foods ..........._ ..... ..__ ...............111.
Figure 5-2 Overlap of domains for pastimes ..........._ .....___ ...............111
Figure 5-3 Overlap of domains for things ...........__......___ ...............112.
Figure 5-4 Cut off point ........... ..... .._ ...............122..
Figure 5-5 Overlap of foods domain reduced dataset ..........._ ..... ..__ .. ....__ ........2
Figure 5-6 Overlap of pastime domains reduced dataset ................. ................ ......... .124
Figure 5-7 Overlap of things domain Reduced data ................. ...............125........... .
Figure 6-1 Plot of Errors food .............. ...............144....
Figure 6-2 Plot of Errors pastimes ................. ...............145........... ..
Figure 6-3 Plot of Errors things ................. ...............146.......... ..
Figure 6-4 MDS for American foods (stress = 0. 195) ................ ...............149...........
Figure 6-5 MDS for Migrant foods (stress = 0. 104) ................. ...............149...........
Figure 6-6 MDS for Mexican foods (stress = 0.181)............... ...............150
Figure 6-7 MDS for American pastimes (stress = 0.222) ................. ...............150........... .
Figure 6-8 MDS for Migrant pastimes (stress = 0. 140) ....__. ................. ........._... ....5
Figure 6-9 MDS for Mexican pastimes (stress = 0.231) ................. ...._.._ ..................5
Figure 6-10 MDS for American things (stress = 0.087) ................. ...._.._ ............... ...5
Figure 6-11 MDS for Migrant things (stress = 0.084) ........................... ........._... ....5
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
CULTURAL MODELS AMONG TRANSNATIONAL MEXICAN MIGRANTS
Mark C. House
Chair: H. Russell Bernard
In this study I examine the cultural changes of transnational Mexican migrants. The advent
of low cost air transportation and cheap telecommunications technology has changed the nature
of Mexican/American migration. These facts of modern life have allowed migrants to move
between two countries repeatedly. During their migrations they have begun to live between two
cultures, embracing their experience as long term migrants.
Traditional acculturation theory was developed to explain how migrants adapt to a new
culture after their migration. This theory was built on the assumption that the migration was
permanent and migrants would establish themselves in their new society. The new environment
of migration has spawned new theories of migration, including transnational migration. This
theory advances the idea that migrants are maintaining social contracts in two cultures.
Essentially they are living between cultures, yet transnational theory does not speak about how
migrant culture changes. Theories of how migrants change have to be reexamined to determine if
they are applicable to this transnational situation.
My research examines the elements and structure of three cultural domains of migrants' to
understand how they are similar and different from non-migrant Mexicans and Anglo
Americans. This dissertation includes data about their acculturation and transnational behaviors.
My research shows that there is evidence of both Mexican and American cultural
influences among migrants. The knowledge that migrants have of their domains is not related to
their acculturation into American lifestyles. My findings may indicate that migrants today are
fundamentally different than their predecessors. While they may adopt some parts of the
American lifestyle and hold onto parts of their home culture they may not completely settle into
being American, but instead
Advances in technology and air-travel, combined with the changes in the economic and
political landscape of countries worldwide, have changed migrant behavior. The increasing
availability of the Internet, inexpensive phone service and next-day mail allows people to
communicate with their families in almost any part of the world and with little or no expense or
delay. In the past, a trip home required a maj or commitment of time and money. Today, with cut-
throat competition in low-cost air transport, migrants in the U.S. can now fly home for a long
weekend to Mexico or the Caribbean (or from Europe to Turkey or North Africa). This ease of
travel and communication has created an environment where migrants no longer have to sever
ties in their home country to come to the U.S. and they no longer have to give up their culture to
become an American.
In response to this changing environment of migration, the concept of 'transnationalism'
came to prominence in the early 1990s for describing people who maintain lives in both their
home and host countries (Portes 2001). Transnational theory focused on why people migrate
repeatedly between cultures, how such migrants influence their sending and receiving
communities and what the larger effects of these migration patterns may be. One area that
theories of transnationalism have not focused on is how people change as a result of constantly
moving between countries. This issue has been left to theories and models of assimilation and
acculturation. However assimilation and acculturation are based on the assumption that migrants
are moving from one social or cultural system to another, and involves the replacement of the
home culture with the host culture. These assumptions are brought into question when migrants
are involved in both cultures to a larger extent. The research described in this dissertation
combines transnational and acculturation theory to present specific patterns on aspects of cultural
change during repeated migrations of Mexicans to the United States.
To examine the changes in migrant culture, very systematic and Eine-grained tools are
needed. These tools are available in Cognitive anthropology and related disciplines to examine
the interaction between human thought and our social and cultural world. Cognitive
anthropologists and psychologists developed the concepts of domains, prototypes, schemas,
taxonomies, and cultural models. These concepts give us an understanding of the how people
organize and use knowledge and have been operationalized over the years for measuring
culture--or at least components of culture-systematically. I use the idea of a cultural domain as
a starting point from which to measure changes in people's thought and behavior that result from
their back-and-forth migrations. Data for cultural domains can be collected reliably and with
sufficient detail for systematic comparison across groups. Using these kinds of data, I examine
how the ideas of transnationalism and acculturation are developed in migrants' material lifestyle.
In addition to providing Eine-grained data on components of culture, I systematically study
how cultural domains function. Most of the research in cognitive anthropology has been with
relatively homogeneous cultural groups, or with heterogeneous groups in relatively stable
societies. In the examination of migrant culture, I show how cultural domains can be influenced
by more than one culture. I begin by examining the motivation for the study and the tools
Why Study Migration?
There has been a surge in migration to the U. S. unmatched since the late 19th and early
20th centuries. Unlike the earlier waves of migrants who came mostly from European countries,
this time our closest neighbor, Mexico, is sending the majority of migrants. The U.S. Census
(Census 2000) estimates that there were 41.3 million Hispanics in the United States in July 2004,
out of a total population of 293.7 million an increase of 84% from the 22.4 million Hispanics
estimated to have been in the U. S. in 1990. Mexican immigration to the U. S. is estimated at 3.5
to 5 million people per decade until at least 2030 and, regardless of economic conditions, to
double by 2030O (Consej o Nacional de Poblacion November 2001). Velez-Ibafiez (2004)
estimates that by 2100, 60% of the population of the U. S., both native and foreign bomn, will be
of Mexican origin.
Just as the sending counties differ, the patterns of migration also differ from that of earlier
generations. Migrants from Europe in the late 19th and early 20th centuries came from distant
lands at a time when the infrastructure made international movement difficult. Today,
international air fares have fallen, making it easy for migrants to come and go, especially
between countries that share a border, like the United States and Mexico. Additionally, the
Immigration and Reform Act of 1986 (IRCA) legalized nearly 3 million illegal migrants who had
been working in the U.S. This removed the risk to previously illegal migrants of returning to
Mexico for visits since they could legally re-enter the United States (Basch, et al. 1994;
Irzigsohn and Saucedo 2002). Besides having the ability to travel easily, migrants from Mexico
and the Caribbean are in continual communication (using inexpensive phone cards) with
relatives and friends in both their communities of origin and their communities of destination
(Basch, et al. 1994; Itzigsohn, et al. 1999; Landolt 2001).
Growing up on a farm in western Kansas, I saw migrants arrive in the spring, when
agricultural employers had work for them, stay through the summer and fall, and return to
Mexico for a short time during the winter. The concept of transnationalism goes beyond taking
the occasional trip home for the holidays and defines these migrants as people who have
developed a lifestyle of migration that is qualitatively different from that of traditional migrants.
Transnationalism involves migrants creating and maintaining two sets of social relationships, one
in their home country and one in their host country (Basch, et al. 1994; Irzigsohn and Saucedo
2002) and building "social fields that link together the migrants' country of origin and their
country of settlement" (Glick Sheiller 1995: 171). The fraction of transnational migrants has not
been measured, but estimates range from 20% to 70% of all migrants (Cohen, et al. 2003;
Irzigsohn and Saucedo 2002; Levitt, et al. 2003).
With changes in the patterns of migration, researchers are questioning some long-
standing ideas of migration. In the mid-20th century, social scientists used the term
'acculturation' to describe the similarities they observed in cultures located near each other. In a
classic paper, Redfield et al. (1936) defined acculturation as, "those phenomena which result
when groups of individuals having different cultures come into continuous first-hand contact,
with subsequent changes in the original cultural patterns of either or both groups."
There have been some issues with this idea of acculturation as it was defined by Redfield.
This definition takes for granted that acculturation is a social or group-level process. It also
assumes that: there is long term contact between the groups; each culture has a defining
character; and that changes can occur in both the sending and receiving societies (Trimble 2002).
With the shifting nature of migration, researchers developed new theories about how
migrants adapt after a move. However, these theories do not provide specific guidelines for
understanding the changes that occur in migrants once they have begun living a transnational
Why Use Cultural Domains?
For many cognitive anthropologists, culture is "whatever it is one has to know or believe in
order to operate in a manner acceptable to its members" (Goodenough 1957: 167). This
definition came out of a crisis in the reliability of representation of cultures and focuses our
attention away from the material artifacts produced by a culture and onto the knowledge that is
required for their production. Early efforts in cognitive anthropology focused on systematically
understanding various semantic domains in different cultures using feature analysis (Conklin
1954; Goodenough 1956; Lounsbury 1964b). Classic feature analysis fell out of fashion for good
reason (see Burling 1964); but the cognitive sciences turned to more complex cognitive
structures with the development of prototype theory (Berlin and Kay 1969; Lounsbury 1964a;
Rosch 1978b). Prototypes are a cognitive representation of an obj ect that aid in its classification
using typical features (Rosch 1978b). Prototypes were generalized into the theory of schemas by
allowing variation from the environment to be fitted within open slots in the model whereas in
prototypes these slots were filled with typical instantiations (Bartlett 1932; Langacker 1987;
Rumelhart 1980; Wallace 1970). Schemas are "conceptual abstractions that mediate between
stimuli received by the sense organs and behavioral responses (Casson 1983: 430)." These
mental entities are used to classify obj ects and to understand spatial relationships and events; in
other words they organize experience (Casson 1983; Mandler 1984; Rice 1980).
Schemas are found at different levels of generality. Universal schemas are shared by all of
humankind while idiosyncratic schemas are the products of an individual (Rice 1980; Shore
1996). Cultural schemas or models provide a midpoint between these poles (Rice 1980) and are
shared to some degree with other people in a society (D'Andrade 1987). These models define the
elements of a particular domain and how these elements are structured. They structure our
understanding, expectations and inferences about the world as well as the decisions we make
(Quinn and Holland 1987). Cultural models have been developed for American gender
terminology (Holland and Skinner 1987), buying behavior (Rumelhart 1980), the western idea of
mind (D'Andrade 1987), American marriage (Quinn 1987), anger (Lakoff and Kovecses 1987),
and the way that heating thermostats work (Kempton 1987), to name a few. Similar to cultural
models, cultural domains allow the researcher to examine a very small part of a culture, but in
great detail. However cultural domains are created with systematically collected data and can be
analyzed with statistical tools. These two characteristics are necessary to especially helpful in
understanding how culture changes as a result of migration.
By examining the cultural models of migrants I also attempt to expand the usefulness of
cultural models by examining how they are affected by multiple cultural influences. Romney and
Moore (1998) suggest that examining the structure of cultural domains in bilinguals and
comparing them to the domains of monolingual people would significantly advance our
understanding of culture as shared cognitive representations of semantic structures. Similarly,
Schrauf (2002) argues that instead of comparing separate cultural groups, an alternative is to
examine the cognitive domains of immigrants. Because immigrants would have two competing
cultural influences the investigation would highlight the differences between two cultures. The
research reported here answers these calls by examining a sample of cultural models from
Mexicans who move regularly between their home in around the city of Guadalaj ara in the State
of Jalisco, Mexico and Goodland, Kansas and comparing these models with those of non-
migrants living in the US and non-migrants living in Mexico.
Chapter one gives a broad overview of the topics in question and ties the methods of
cognitive anthropology with the topic of migration. Chapter two provides a historical and
theoretical context for this research: a brief history of U. S./Mexico migration, followed by a
general review of migration theory and related factors. The chapter concludes with a discussion
of theories of acculturation and transnationalism. Chapter three reviews the history of cognitive
anthropology to provide an understanding of why cultural models are useful in examining
migrant culture. Chapter four explains the research design, the sampling strategy and the data
collection methods used for this proj ect. Chapter five presents the analysis of the elements of the
domains. Chapter six is an analysis of the structure of the domains and the relationship between
migrants' cultural knowledge and their transnational behavior and acculturation. Chapter seven is
a discussion of the results and future directions for the research begun here.
This chapter provides a background on the history and theory of migration between the
U. S. and Mexico. Any effort to recount the history of Mexican migration to the U. S. must
consider the policies that have shaped this movement. While not always achieving the goals they
are written for, immigration policies have a profound affect on the number of migrants entering
the country, how they enter the country and where they move to after entry (Massey and
Espinosa 1997). To give a context for the research presented here, the first section presents a
short history of U. S. migration law and Mexican migration to the U. S.
The research on migration is vast and is not easily summarized. However, to provide a
theoretical background for this research, the second section examines various theories of why
people migrate, why migration patterns are perpetuated and other issues in migration research,
including the role of racism as either a push or a pull factor; differences between temporary and
permanent migration; the importance of legal status of migrants; the role of remittances for
senders and receivers; and the role of labor organizations in migrants' behavior.
The final section narrows the discussion to theories of post-migration change including
assimilation and transnationalism. Assimilation and acculturation are theories about what
happens to people after they move to a new country and culture. Transnationalism is a collection
of ideas about the changing nature of migration. Applying the idea of acculturation to the
environment proposed by transnationalism, I develop hypotheses about how the culture of
transnational migrants changes during the process of making repeated trips between countries.
Mexican U.S. Migration and Policy
The first immigration law in the U.S. was the Naturalization Act of 1790 which stipulated
that anyone who was a free white person could become a citizen of the United States. Laws in
1862 and 1875 restricted the immigration of specific groups, mainly Orientals, criminals and
prostitutes. Additional restrictions were created in Federal laws enacted in 1882, 1885, 1891,
1903, 1907, and 1917, based on economic, moral and physical criteria (Batchelor 2004).
Before the Mexican Revolution in 1910, Mexican migration was less than 1% of the total
migration to the U.S. (Freeman and Bean 1997b). During this time, migration to the U. S. was
discouraged by Mexico' s system of land tenure, where a small minority owned the vast maj ority
of the land. Those without land made their living through various sharecropping arrangements,
selling homemade crafts and engaging in small scale business operations (Bean, et al. 1997).
This lifestyle required the labor of the entire family, which restricted the ability to of Mexicans
to move abroad.
Incentives for migration during the early 20th century were created in both Mexico and
the U.S. Porfirio Diaz's economic modernization plan included the building of railroads through
the rural areas of Mexico. These railroads were connected to the system of tracks in the
Southwest U.S., making it possible for migrants to move north. The Mexican Revolution of 1910
disrupted the social stability of rural areas through the destruction of crops and trade (Garza and
Szekely 1997). This instability continued with the Cristero Rebellion of 1926, encouraging
further migration. This was especially prominent in the north and west-central states of Mexico
(Garza and Szekely 1997). On the U.S. side at that time, recruiters sought rural Mexican labor
for building railroads, for working in mines and for agricultural work, especially in the
Southwest (Durand, et al. 2001; Fussell 2004). Additionally, World War I cut traditional sources
of labor from Europe further increasing the demand for Mexican labor (Durand, et al. 2001).
While these factors increased Mexican migration during the early 20th century, the total
Mexican migration was not exceptional when compared to migration from other areas (Bean, et
al. 1997). A total of 459,000 Mexican' s registered as migrants to the U. S during the entire
decade of the 1920's. Several European countries contributed equally and Canada had more than
twice this number of migrants to the U. S. during the same period (Bean, et al. 1997). One
feature of Mexican migration that separated it from that of other countries was the seasonal
nature of the migrant work. This encouraged an annual emigration back to Mexico that was
nearly three quarters of the immigration to the U.S. (Bean, et al. 1997). Even at this early stage
in the establishment of Mexican/American migration routes there is evidence that migrants were
making repeated j ourneys between the two countries.
The 1917 Immigration Act increased the fees to enter the country to 8 dollars (135 in
2005 dollars) and imposed a literacy test on all immigrants. However, under pressure from
agricultural producers in the southwest, special consideration was given to Mexico when the
Secretary of Labor waved the literacy test and head tax for Mexican workers. The 1921 Quota
Act created quotas for countries in the Eastern Hemisphere, dictating the number of immigrants
allowed into the U.S. per year from each country. There were however no limits on immigration
from countries in the Western Hemisphere. Countries were also favored in the 1924 National
Origin Quota Act that also excluded them from quotas. This act did however create the Border
The Bracero Program
In 1929, the lack of jobs for American workers created by the Great Depression caused
the U. S. government to limit Mexican migration (Durand, et al. 2001). Deportation of Mexican
workers increased dramatically until World War II (Durand, et al. 2001). In 1942, the Bracero
program created a guest worker program for seasonal agricultural laborers (Durand, et al. 2001;
Fussell 2004). This program gave workers a six month renewable visa to work for approved
growers in the U.S. (Durand, et al. 2001) The Bracero program was beneficial to both the
Mexican and U. S. economies. It provided cheap seasonal work for U. S. farmers and inj ected
cash into the poorer areas of Mexico through remittances. The program was ended in 1964, but
through it more than 4.6 million Mexican workers came to the U.S. (Cornelius 1978).
During the Bracero period, migration increased not only because of the ease with which
migrants could enter the U.S. but also because U.S. wages were many times higher than in
Mexico (Bean, et al. 1997). Migration came especially from the western regions of Mexico
where there were established linkages to the Southwest U.S. and agriculture was more market
oriented creating a greater use for cash (Bean, et al. 1997; Cornelius 1978). During this period
the Mexican government was involved in agrarian reform, begun in 1934, that broke up the
haciendas and created smaller ejidos or communally owned plots. With population growth at the
village level, these plots quickly became too small to be useful and further encouraged rural
peasants to migrate (Garza and Szekely 1997).
Post War Immigration
In 1952, the McCarran-Walter Act was the first act of Congress to consider labor market
shortages directly in U.S. Immigration law. This act favored migrants with skills that were in
short supply in the U.S. and gave special priority to family members of citizens (Freeman and
Bean 1997a). This act did not consider the increasing illegal immigration coming from Mexico.
However in June of 1954, 'Operation Wetback' was launched returning illegal migrants to
Mexico. This action was such a success that illegal migration was insignificant for more than a
decade (Reimers 1982).
The Immigration and Nationality Act was amended in 1965 to eliminate the national
origins quota system, moving U.S. policy away from country-specific visa limits and towards a
system that was blind to the origin of migrants. It also created the first annual cap for migrants
from the Western Hemisphere (Batchelor 2004; Freeman and Bean 1997a). However in 1976,
Congress imposed a limit of 20,000 visas per country in Latin America and eliminated
preferential treatment for migrants from the Western Hemisphere. This law also eliminated the
cap for family members of citizens, creating a massive backlog of applicants from Mexico
(Freeman and Bean 1997a).
The most recent influential reform of immigration law in the U. S came in 1986 in the
form of the Immigration Reform and Control Act (IRCA). IRCA had three components aimed at
reducing the flow of illegal migrants to the U.S. First, the act attempted to eliminate the
attraction of U. S. j obs by sanctioning employers who knowingly hire undocumented migrants.
Employers who repeatedly hire illegal migrants can be penalized with a 10,000 dollar fine and
criminal prosecution. Second, the act gave additional resources to the US border patrol,
including 400 million dollars to hire additional officers. Third, the act gave amnesty to illegal
migrants who could prove that they were in continuous residence in the U.S. after January 1,
1982. A special clause was included giving undocumented farm workers legal status if they
could demonstrate that they worked in the U.S. for at least 90 days during the twelve months
proceeding May 1, 1986 (Durand, et al. 2001; Freeman and Bean 1997a; Massey 1987; Phillips
and Massey 1999).
The effects of the IRCA are still being researched. The amnesty proviso legalized almost
3 million formerly illegal migrants, of whom over 70% were from Mexico (Freeman and Bean
1997a). Some 1.7 million of these migrants demonstrated long term residence and 1.3 million
workers were legalized through the special agricultural workers allowance (Phillips and Massey
1999). Not counting the reduction in illegal migration that resulted from the massive
legalization, the question still remains: has the bill reduced the flow of illegal migrants?
Because of methodological difficulties collecting data on illegal migrants, there are mixed
conclusions on this question. On the one hand, there were fewer arrests of migrants crossing the
borders for the three years following the implementation of the IRCA in 1986 (Freeman and
Bean 1997a). On the other hand, sending communities (communities that send migrants) show
no decline in the number of migrants moving across the border (Donato 1994). Cornelius
(1978), Durand and Massey (1992) Further evidence suggests that employee sanctions may not
be effective at reducing illegal migration as these were in full effect only after three years when
arrests were once again on the increase (Freeman and Bean 1997a).
Other effects of the IRCA include a reduction in wages for illegal migrants (Phillips and
Massey 1999). Employers increasingly used subcontractors for labor after IRCA. These
subcontractors took the risk of hiring illegal migrants away from employers but passed this risk
on to the migrants in the form of reduced wages and poor working conditions (Phillips and
Massey 1999; Taylor, et al. 1996).
At the time when the IRCA was being implemented, there were a number of factors in
Mexico that encouraged both legal and illegal migration to the U. S. Roberts and Latapi (1997)
reviewed these issues and concluded that changes in education, health and housing did not
contribute to increased migration. While these areas did not worsen during that time, they
showed slowed improvement. Roberts and Latapi (1997) did find that increased income
inequality contributed to migration. This resulted when waged employment decreased and people
were forced to turn to self-employment (Roberts and Latapi 1997). With the end of subsidies for
urban consumers, urban areas stagnated and could no longer attract rural migrants (Roberts and
Latapi 1997). At the same time, crop prices were no longer subsidized and land was less
available, creating additional incentive for rural workers to migrate (Latapi, et al. 1996).
The 1990 Immigration Act raised the annual ceiling for certain categories of immigrants.
This act attempted to attract skilled labor for U. S. businesses. The 1996 Immigration Reform
Law and Welfare Reform Law attempted to curb illegal migration by doubling the number of
border patrol agents and making it impossible for undocumented immigrants to get Social
Security benefits as well as most federal, state and local assistance (Batchelor 2004).
Additionally, legal immigrants were excluded from most federal programs for 5 years.
Mexican/American migration is somewhat unique in the world because of the sharp
contrast in living standards between the two countries and the fact that they share an extensive
stretch of almost unguarded border. These two factors have created an exchange of labor, money
and products that both countries have come to rely on. The policies of both countries have
shifted through time from encouraging migration to attempting to deter it altogether. However, I
think it is uncontroversial to say that people (mostly Mexicans) will continue to travel between
the countries and blend their cultures for many years to come.
Theories of Why People Migrate
Migration is not only situated in a political environment, it also has a significant
theoretical base that has developed over decades of researchers examining the phenomenon.
Economics, sociology, anthropology, history and political science all have a large number of
researchers dedicated to the study of migrants and their movement around the globe.
While there are differences in each discipline's study of migration, some basic questions have
guided all the research, such as: why do people migrate; why do people return home; what
happens after migration; how do migrants adjust to a new society; and what are the effects of
migration on sending and receiving countries? Different disciplines have not always agreed on
the methods or even the level of analysis that should be used to investigate these questions.
However because of this diversity of perspectives, a large variety of theories have been
developed to deal with the questions of migration.
Early Equilibrium Theory
According to Portes (1976), the earliest work on migration in anthropology derives from
the unilinear evolutionary theories of the 19th and early 20th centuries. These theories placed
societies in various continua-Tonnies' Gemeinschaft and Gesellschaft societies, Durkheim's
mechanical vs. Organic solidarity societies, and so on. Redfield (1941) saw a folk-urban
continuum, with societies in rural, undeveloped, agricultural areas at one end and large,
developed cities at the other. Applying this model to migration, he assumed that the migrant
would assimilate the modem innovations they encounter in urban society and return these ideas
to their rural regions (Keamey 1986; Miracle and Berry 1970). These innovations would then
lead to the modernization of the rural sending areas.
During the 1960s and 70s, researchers noted that migrants were not acting according to
the predictions of these early models. Gmelch (1980; Swanson 1979) reviewed research on
return migration and found that return migrants generally did not bring skills that were useful in
their home economy nor did they invest their savings in the community; instead they used their
earnings for housing and to buy consumer goods. Agricultural workers might buy land with their
savings but would lease it out rather than farm it and provide themselves with an income stream
(Wiest 1979). Another questionable assumption was that because rural economies had a surplus
of labor, they should spontaneously export it. History has shown that the export of labor is rarely
spontaneous or even consensual (Portes 1978). Instead of creating equilibrium between labor
and capital, as the model predicted, migration was escalating the disparities between rural and
urban areas (Amin 1974).
Neoclassical Economic Theory
Drawing from the ideas found in modernization theories, neoclassical models from
economics used the idea of two types of societies to create models of migration based on
differential amounts of labor and capital in different areas (Kearney 1986; Massey 1987; Wood
1982). Areas with high wages would attract migrants from areas of low wages. Inversely places
with low capital would have higher possible rates of return and thus would attract investment
capital. Additional money would be sent from areas with high capital through the remittances of
migrants sent to their family and friends. These two flows would eventually create equilibrium
between two areas through migration (Massey, et al. 1993). The neoclassical model of migration
was a model of individual choice. This model stated that the decision to migrate was a cost
benefit analysis that considered several key factors (Todaro and Maruszko 1987). Research
using this theory has found that migration is positively related to a high wage ratio at the
destination, a large labor force at the origin and destination, and negatively related to high wages
at the migrant' s home and the distance between the origin and destination cities (Greenwood, et
al. 1981; King 1978; Rogers 1968).
Other work indicates that Mexican migration is negatively related to the farm wages in
Mexico and agricultural productivity in Mexico and positively related to farm wages in the US
(Frisbie 1975). In accordance with the neoclassical model, Jenkins (1977) found that the Mexico
- US wage differential was positively related to Mexican migration. Further refinement of this
model indicates that it is not the real wage differential between sending and receiving
communities but the expected earnings gap that is significant (Taylor 1987; Todaro and
Maruszko 1987). Unemployment in Mexico has also been found to be positively related to
Mexican out-migration (Blejer, et al. 1978; White, et al. 1990). There is considerable research
supporting neoclassical theory. However, Massey and Espinosa (1997) found that while the wage
differential is weakly associated with the odds of taking a first trip to the United States, it is not
related to making additional trips or return home trips.
Dual Labor Market Theory
Dual or segmented labor market theory is a macro level theory that is similar to
neoclassical theory with one exception. While dual labor market theory agrees that migration is a
rational decision made by individuals responding to labor markets around the world, it sees
migration as a result of a segmented or dual labor market that is inherent in the economic
structure of advanced industrial societies (Piore 1979). Instead of a continuous labor market as
assumed by neoclassical economics, the labor market is divided into two sectors: a primary
sector that provides good wages, working conditions and steady employment, and a secondary
sector with opposite traits (Dickens and Lang 1988). In developed societies, there is a shortage of
workers who are willing to work in the secondary sector, so employers are forced to tumn to
migrant labor (Piore 1979).
This theory was prominent in economic research in the mid 1970s, but testing the theory
yielded mixed results (see (Massey, et al. 1994). The theory reemerged in the late 1980s with
some improvements in the model that has considerably improved its fit with empirical data
(Dickens and Lang 1988). Applying this theory to immigration, immigrant workers who take
jobs in the secondary sector would be expected to have lower returns for their education, skills
and other work experience than native workers (Massey, et al. 1994). Using cross-sectional data,
Chiswick (1984) has found evidence contradicting these predictions. However other studies have
found that education, experience and j ob skills in secondary sector employment are negatively
related to migration (Lindstrom and Massey 1994; Portes and Bach 1985). These studies have
also found that English language ability and time in the U.S. were both positively related to the
ability to work in the primary sector or to gain higher earnings in the secondary sector.
New Economics of Migration
The 'new economics of migration' approach and the neoclassical economic view are both
micro-level decision models that rely on rational choice of the deciding unit. However a key
insight of the new economics of migration approach is that migration decisions are not made by a
single person, instead it is the family or household where the decision is made (Wood 1982).
Because the decision lies within a small group of people, there are other considerations in the
decision to migrate (Stark and Bloom 1985). These factors can include the likelihood of crop
failure, risks of falling market prices, unemployment by some members of the household or the
desire to decrease the household's relative deprivation (Massey, et al. 1993).
Even if the migrant does not increase the absolute income of the family, they provide an
important way to mediate these risks by either sending money home or returning home with
earnings. Relative income, or the income in comparison to the migrants peers, may be more
important to migrants than absolute income (Massey and Espinosa 1997; Stark 1991). Another
flaw found in the neoclassical model that is highlighted by the 'new economics of migration'
literature is that it assumes that people will migrate on a permanent basis, despite the fact that
regular temporary migration has been documented for a number of countries (see Massey, et al.
1994). There is also evidence that when migrants do return, they spend their earnings on income
producing assets like additional land, livestock or equipment (Fletcher and Taylor 1992; Taylor
and Wyatt 1993).
Raul Prebisch, head of the United Nations Economic Commission on Latin America.
Prebisch in the late 1940s, saw that, contrary to neoclassical predictions, economic growth did
not benefit some nations as their wealth continued to decrease while the wealth of others
increased. In The Economic Development of Latin America and' its Principal Problems (1950),
Prebish theorized about these differences and produced what came to be known as dependency
theory. The Singer-Prebisch thesis (so named because Hans Singer, another United Nations
employee, developed the same idea at the same time) stated that poor counties export raw
materials used by rich countries that would then manufacture these materials into finished goods
and resell them back to poorer countries (Singer and Chen 1998). However, because of unfair
trade policies maintained by the rich countries, the poor countries would never earn enough from
their exports to pay for their imports.
Ferrara (1996) describes three assumptions that were generally shared by dependency
theories. First, these rich and poor countries were renamed as core and periphery (Wallerstein
1974), metropolitan and satellite or dominant and dependent. States with low per capital GNP and
that relied on the export of raw materials were classified into the second category of the scheme.
Second, both types of states are connected through historical, dynamic and ongoing relationships
that reinforce the dominant/dependent connection. This relationship was created by the policies
of trade that favored the core nations. Companies in core nations would retain profits through
developed unions, while companies in peripheral nations would not have this advantage. Third,
external forces like multinational corporations, international markets, foreign aid and any other
means by which dominant countries can further their influence are the primary factors in the
relationship between the two types of states. Instead of developing the rural regions of the
world, dependency theory saw the movement of labor to core areas as exploitation that benefited
the developed parts of the world (Wood 1982). Kearney (1986) notes that while dependency
theory has focused research onto the extraction of human resources via migration it, "does not
provide anthropologists with a general theoretical model capable of generating many local level
research problems on migration emanating from rural communities" (p. 340).
World Systems Theory
World systems theory has continued the ideas developed in dependency theory
hypothesizing a single worldwide system of labor and commodities that are exchanged among
different regions (Wallerstein 1974). Instead of having a bifurcated system of folk and urban
areas, the regions are classified into a periphery, semi-periphery, or core areas. Migration occurs
because of the penetration of capitalist economies into non-capitalist societies that is a product of
globalization (Massey, et al. 1993). Disruptions to the existing economy result from the
introduction of capitalism in various forms, including: land reform, extraction of raw materials
and the labor demands of new factories. These factors can cause large numbers of people to lose
their traditional livelihood and become prone to migrate (Chayanov 1966; Massey, et al. 1994).
This system originated as a result of colonization of Asia, Latin America and Africa that allowed
Europe to have access to the markets and natural resources worldwide. More recently Europe
and the U.S. have used diplomacy and their military to develop and protect their investments,
resources and allies (Rumbaut 1991). This theory has been used in some anthropological
research on migration (Rhoades 1978; Wiest 1973) but has proven difficult to test directly.
However there are some indicators linking economic development in a region to out-migration
(Hatton and Williamson 1994; Roberts 1982) Similarly, direct investment by the U.S into
foreign economies is a strong indicator of market penetration (Sassen 1988) and is highly related
to the annual rate of migration to the U.S. (Ricketts 1987).
Unlike world systems theory, articulation theory does not see a single global capitalist
system of societies (Kearney 1986). It does acknowledge the division of core and peripheral
societies but peripheral communities can have qualitatively different forms of economy other
than capitalism and these other economies can coexists along side of capitalism to some degree.
Another difference from World Systems theory is that articulation theory focuses on production
instead of circulation as the primary means for the transfer of surplus to capitalist economies.
Research in this tradition focuses on the creation of excess labor it transfers to capitalist societies
as the prime factor in migrant labor. Articulation theory states that the economy that produces
migrants is destroyed because the labor is drawn off to work in the capitalist economies
(Meillassoux 1981). As this labor is drawn a temporary migration attempts to save the domestic
non-capitalist economy, which in turn creates a permanent state where migration is the norm and
not a transition to a capitalist economy (Meillassoux 1981; Rey 1973).
Summary of Factors in the Creation of Migration Streams
There are many variables that account for the ebb and flow of migration from
Mexico(Massey, et al. 1994) but there is no single theory to account for the phenomenon. An
excellent attempt at understanding the strengths and weaknesses of these different theories was
made by Massey and Espinosa (1997) who use one dataset to test several theories against one
another. The results provide and excellent summary of the factors influencing migration. They
indicate that people with social ties to other migrants are more likely to migrate initially. One
assumption of neoclassical economic theory is that people with physical capital would be more
likely to migrate. However their results show that people with homes, land and a business are
less likely to migrate and especially are less likely to be migrating illegally. Nevertheless,
neoclassical economic theory does show us that the likelihood of undocumented migration is
related positively to the U.S./Mexico ratio of expected wages, although this is not the best
predictor of migration. Completing high school (preparatoria) in Mexico lowers the odds of
migration while the presence of a bank in a person's home town raises these odds, supposedly by
providing a means to safely accumulate the earnings of migrants (Massey, et al. 1994).
Congruent with predictions of the new economy of migration and with world systems theory,
high levels of development in a community as well as high wages increase the likelihood of
undocumented migration as does the larger proportion of workers earning twice the minimum
wage and the proportion of women working in manufacturing j obs (Massey, et al. 1994). Massey
et al. support segmented labor market theory's assumption that migration follows U.S. labor
demand. World systems theory states that migration is an effect of capital penetration; however
the rate of growth in direct foreign investment decreases the odds of migration in their study
(Massey, et al. 1994). The supply of visas is negatively related to the number of undocumented
migrants, but efforts at controlling undocumented workers are not effective. Rather than
discouraging employment of undocumented workers, sanctions against employers and an
increase in resources for border restrictions increase the likelihood of taking an initial trip to the
U. S. Furthermore, the legalization that were a part of IRCA have increased migration by
increasing the network ties available to non-migrants. Massey and Espinosa (1997) sum up the
contribution of many different theories.
In accordance with the neoclassical model, the likelihood of illegal migration is
positively related to the U.S.-Mexico wage differential. Being related to a migrant
sharply boosts the odds of taking a first U. S. trip, which is consistent with social capital
theory. In keeping with the new economics of migration, high real interest rates increase
the probability of undocumented movement. Following segmented labor market theory,
undocumented migration is linked to the growth of U. S. employment. And, consistent
with world systems theory,..., the odds of undocumented migration are greatest in
dynamic, developing communities, not stagnant area with low wages and marginal levels
of industrialization. (p. 964 965)
Why Does Migration Continue?
Besides examining factors that initiate migration, there is substantial research on factors
that perpetuate additional migration. These factors may either increase the likelihood that the
migrant will make repeated trips or they may reduce the costs and risks for family and friends to
also migrate (Massey, et al. 1994). Findings that support the former have been dubbed
'cumulative causation' (Massey 1990) and show that the probability of making multiple trips is
increased with migrant experience in the new country (DeJong, et al. 1983; Massey 1987; Root
and DeJong 1991; Taylor 1987) Research on the latter, reducing the costs and risks to migration,
has used social networks to understand the factors involved.
When migrants have made the trip abroad, they return with a vast amount of knowledge
and contacts that can be useful to their friends and family. These social networks increase the
likelihood of migration by decreasing the risk and costs associated with moving to a new country
and starting a new job (Choldin 1973; Ho 1993; Massey and Espana 1987). Cost is reduced as
the number of migrants increase. Once the first migrants from an area are established, they
provide assistance and knowledge that eases the move for future migrants (Wilson 1998). Risk is
also reduced with the increased number of migrants until at some point the household relies on
the income from migrants as part of their survival. Similar results have been noted by a number
of researchers who have found that social ties to someone in the U.S. is the best predictor of
migration (Chavez 1988; Fjellman and Gladwin 1985; Massey 1987; Massey and Espana 1987).
Additional work with migrant networks has yielded some interesting results. Cohen et al.
(2003) found that people who know more individuals with migration experience are more likely
to be migrants themselves. Guarnizo et al. (2003) found that network size, but not network
composition, or who was in the network, was positively related to transnationalism migration.
Working in Mexico, Curran and Rivero-Fuentes (2003) found that male networks are more
important for moves to the U.S., while networks composed of female members decreased the
odds of a male migrating. Female networks did however increase the likelihood of female
migration and were especially important in migration within Mexico. Kanaiaupuni (2000) found
similar results suggesting that high female employment at the destination encourages female but
not male migration. The size of the population in the receiving country has been found to be
critical in attracting migrants (Dunlevy 1991) as has the network density (Winters, et al. 2001).
Winters, et al. (2001) also found that as migrant networks become established, community
networks can be substituted for familial networks. In an examination of the types of relationships
that compose the network, the degree of kinship plays an important part in the decision to
migrate within Mexico as well as to the U.S. (Davis, et al. 202).
When migration back and forth has become self sustaining over time, a culture of
migration comes into existence. This culture of migration creates social norms that encourage
migration in a number of ways. Massey et al. (1994) have divided these norms into three
categories: first migrants are able to engage in the consumption of consumer products that is
beyond the means of non-migrants (Fletcher and Taylor 1992; Georges 1990; Massey 1987;
Wiest 1979; Wiest 1973); second migration becomes a rite of passage for young men in the
community (Georges 1990; Reichert 1982; Rouse 1992); and third migration influences gender
relationships within the family as women migrate to societies with fewer constraints on female
autonomy (Georges 1990; Grasmuck and Pessar 1991).
Other Factors in Migration
There are a variety of other factors that are significant in the study of migration. Some
may not contribute directly to the creation of or maintenance of migration streams but are no less
important. A few of these issues are examined below.
Racism is certainly a factor when people move into a new culture. It is compounded
when the phenotypes of the sending and receiving societies are different and the racial categories
in the two societies are different as well--as is the case for the U.S. and Mexico (Alonso 2004).
However the literature linking migration and racism in Mexico is very sparse. The investigation
of racism in Mexico has included the relationship between class and ethnicity (Alonso 2004;
Lewis 2000; Stephen 1991) and the relationship between native ethnicity and political
movements (Nagengast and Kearney 1990).
On the other side of the border the racism against Mexicans provides historical
perspective. Menchaca (1993) examines how the legal system of the U. S. discriminated against
Mexicans from 1848-1947. She concluded that Mexicans who were of indigenous descent were
more rigorously discriminated against. She also found that phenotype played a large part in the
degree of discrimination (Menchaca 1993). Velez-Ibanez (1996), working in the Mexican
American community in the Southwest, shows how this population reacts to the racism they
experience in the U.S. Unfortunately there is little work connecting migration and racism in
Mexico or the U.S. I suspect that this may result from the lack of discernment by Americans
between Mexicans and Mexican Americans. Both groups are likely to be discriminated against
based on their race with little acknowledgement of their citizenship.
Temporary Versus Permanent Migrants
One question within migration work is what makes a migrant stay in the U. S? Massey
and Espinosa (1997) found that the likelihood of returning to Mexico was increased by being
married, by the number of prior trips to the U. S., by owning land or a home, by living in a
community with a bank, and by increases in Mexican inflation and interest rates. The likelihood
of returning to Mexico was decreased by increases in education, by increased experience in the
U. S., by the duration of the trip to the U. S., by holding a skilled or unskilled j ob, by having
children in the U. S. and by increased availability of visas (Massey and Espinosa 1997). While
the availability of visas may reduce the likelihood of returning to Mexico Durand, et al. (2001)
found that the massive increase in legalization in the late 1980's actually increased the return
Canales (2003) examined the macro influences of Mexico' s economic policies and
business practices within the U.S. on Mexican migration. He concludes that the neoliberal
policies in Mexico have reduced j ob security as well as the overall number of jobs and total
wages. He also found that the labor market in the U.S. has split into high paying professional
jobs and low paying unskilled labor. This dual labor market has also been found by other
researchers (Piore 1979). Canales argues that Mexican labor has moved out of the traditional
agricultural work towards other types of unskilled labor that provide year round employment,
thus increasing the number of permanent migrants to the U. S.
Massey and Espinosa (1997) examined variables that influence the likelihood of taking a
first trip to the US legally or illegally. Both legal and illegal migrants were influenced by having
a parent or sibling who migrated, and living in an agrarian economy. Legal migrations increased
with an increase in the number of migrants in the community, an increase in the number of
people earning twice the minimum wage in the community, an increase in the number of females
employed in manufacturing positions in the community, and with having an ejido established.
These variables were negatively related to illegal migration. A number of other variables were
related to legal migration but not illegal migration, including: Age squared (-), marital status (-),
owning land (+), owning a home (-) or business (-), having attended a preparatory school (-),
having a bank located in the community (+), the Mexican inflation rate (-), U.S. employment
growth (+), growth in foreign investment (-), Mexican interest rates (+), and the availability of
visas (-) (Massey and Espinosa 1997).
For both illegal and legal migrants the factors that influence a return trip to Mexico
include: marital status, education, cumulative U.S. experience, duration of trip, holding an
unskilled urban j ob, number of U. S. migrant children and owning land (Massey and Espinosa
1997). Owning a home, living in a community in Mexico with a bank, the availability of visas,
and the number of self employed in the sending community were significant factors for legal
migrants but not illegal migrants to return home. Illegal migrants were influenced by a different
set of factors to return home, these include: not being a wife of a U.S. migrant, not having U. S.
born children, owning land in Mexico, living in a community in Mexico with a preparatory
school (the equivalent of a high school in the U.S.), having paved road and not an ejido,
increased risk of apprehension and the enactment of employer sanctions (Massey and Espinosa
While not directly influencing the creation or maintenance of migration flows, the study
of remittances from migrants to their home country is certainly pertinent given that these totaled
more than 13.3 billion dollars in 2003 up from 10 billion in 2002 and expected to rise annually
(Coronado 2004). Electronic transfers were the most popular method for sending money to
Mexico with 85.8% of the money sent this way. Money orders followed with 12.2% and cash or
material goods had 1.9% of the total money sent to Mexico (Coronado 2004). The importance of
remittances lies in the fact that migrants can eamn the wages of high-income countries and in
sending some of this income home, they can spend the money at the prices of low-income
There is some disagreement about whether remittances have a positive or negative impact
on origin countries. At the household level, Itzigsohn (1995) found that low and middle-low
income families have a higher overall income if they receive remittances. This point may sound
simple, but it shows that the cost of sending a migrant to work abroad is less than the income
received from them (Keely and Tran 1989). These remittances also allowed some family
members to stay out of the workforce or at least to avoid the most menial j obs and work in
riskier employment (Itzigsohn 1995).
The question can also be asked at the community level. Are the communities that send
migrants to work abroad better off because of the remittances? This question is more complex
because not everyone in the community may send a family member abroad. Jones (1998) found
that the inequality between family incomes was influenced by the level of migration found in the
community. He found that remittances decrease the separation of the low income families and
high income families when migration is new. However when migration is more settled within the
community the findings were reversed and income inequality increased. Besides affecting the
spread of a community' income distribution, Conway and Cohen (1998) argue that the
significance of remittances should be measured by their impact on community investment,
household decision making and community structure.
Labor Organization of Migrants
While examples of the labor organization among migrants may be easy to point out,
documentation is more difficult to find. A typical organization is described as mostly migrant
men, recruited by a crew boss to work as harvesters in the agricultural sector (Nelkin 1970).
These bosses recruited migrants on street corners or in bar sweeps, with many migrants having
no intentions of working in a crew until they are recruited (Nelkin 1970). The crew leader is key
and serves as "contractor, recruiter, camp manager, work supervisor, policeman and banker"
(Nelkin 1970: 476). These bosses use their knowledge of the language, work site and
surrounding area to maintain their control over the workers, who often know little about the
environment beyond their j ob. Zlolniski (1994) argues that this type of organization developed
into the extensive use of subcontractors by formal sector firms. Using these bosses as
subcontractors helped reduce the labor cost and overhead costs of maintaining workers within
the formal sector. Additionally these crews provided the formal economy with an increasingly
flexible work force.
Personal experience in Kansas shows a less organized labor structure. Many migrants
will use personal contacts to implement their migration to the U.S. However these contacts do
not maintain the control over the migrant after arrival. I would argue that this occurs because
agricultural work in the Midwest is plentiful and farmers have a long history of working with
Post Migration Changes
This section focuses on the question of the adaptation of migrants to their new society.
Assimilation theory focuses on how migrants give up the customs of their homeland and adopt
the customs of their new culture. Transnationalism points out that the nature of migration has
changed dramatically from the time when assimilation theory was developed. Recall that
advances in technology and air-travel, combined with the changes in the economic and political
landscape of countries worldwide, have changed how migrants behave. The increasing
availability of the Internet, inexpensive phone service and next-day mail allows people to
communicate with their families in almost any part of the world and with little or no expense or
delay. In the past, a trip home required a maj or commitment of time and money. Today, with cut-
throat competition in low-cost air transport, some labor migrants in the U.S. can fly home for a
long weekend to Mexico or the Caribbean (or from Europe to Turkey or North Africa). In short,
migrants no longer have to sever ties in their home country to come to the U.S. and they no
longer have to give up their culture to become Americans.
The changing behavior of migrants has prompted some researchers to reexamine the
focus of their research and to look at how migrants remain connected to both societies. I focus
next on research about how migrants change after settling out or becoming assimilated into their
new culture (acculturation theory) or during continual migration (transnational theory).
Assimilation and Acculturation
The original model of acculturation and assimilation is due to Robert Park, at the
University of Chicago in the years just before World War I. Park developed a three-stage model
that involved contact, accommodation, and, Einally, assimilation. Redfield et al. (1936) defined
acculturation as, "those phenomena which result when groups of individuals having different
cultures come into continuous first-hand contact, with subsequent changes in the original cultural
patterns of either or both groups." This definition takes for granted that acculturation is a social,
or group-level process. It also assumes that: there is long term contact between the groups; each
culture has a defining character; and changes can occur in both the sending and receiving
societies (Trimble 2002).
Spiro (1955), by contrast, asked whether the acculturation process occurs at both the
individual and group level. Most researchers in migration have worked at one level of analysis -
either group or individual though some have examined how the individual acculturation
process is affected by group acculturation (Eaton 1952; Spiro 1995). A more controversial
question is whether or not acculturation has direction. The original definition allows for the
transfer of cultural traits to both cultures. Early scholars of migration argued that one culture is
always dominant in the exchange of cultural traits (Parsons 1936). While this debate was never
resolved, later scholars simply assumed that acculturation was unidirectional (Devereux and
Loeb 1943; Spiro 1955). Additional work on acculturation identified variations in the process:
migrants could accept one culture's patterns of behavior; they could selectively accept some of
the behaviors of the new culture; or they could create novel behaviors in reaction to their new
culture (Barnett, et al. 1953). Redfield's definition was reworked to explicate some of these
changes and in 1953, the Social Science Research Council defined acculturation as: "culture
change that is initiated by the conjunction of two or more autonomous cultural systems.... Its
dynamic can be seen as the selective adaptation of value systems, the processes of integration
and differentiation, the generation of developmental sequences, and the operation of role
determinant and personality factors" (Barnett, et al. 1953).
Study of the acculturation process fell out of favor in anthropology, while assimilation theory
continued to develop in sociology and social psychology. Gordon (1964) defined assimilation as
the process where people of diverse cultures achieve a unity that is sufficient to allow
coexistence. He argued that each stage of assimilation could develop to a different degree. The
following list are the stages of assimilation:
1) Acculturation Change of cultural patterns to those of host society
2) Structural assimilation Large-scale entrance into cliques, clubs, and institutions of host
society, on primary group level.
3) Marital assimilation Large-scale intermarriage
4) Identificational assimilation Development of sense of people-hood based exclusively on
5) Attitude receptional assimilation Absence of prejudice
6) Behavior receptional assimilation Absence of discrimination
7) Civic assimilation Absence of value and power conflict
Under Gordon's conceptualization, acculturation became an explicit stage of
assimilation. Further refinement of these concepts led researchers to argue that assimilation and
acculturation differ primarily because assimilation requires that the dominant culture accepts the
out group (Teske and Nelson 1974). While acculturation can occur in migrant groups through
interaction with a culture, their assimilation cannot occur without being accepted by the
community (Hirsch 1942; Spiro 1955).
The classic, unidimensional definition of assimilation states that migrants adopt the
norms and habits of the dominant society to which they have moved, eventually dropping their
previous way of life. The change is directed by the dominant society imposing their cultural
habits on the migrant. Researchers across the social sciences examined the variables influencing
the assimilation process. Trimble argues that the personal situation of the migrant may determine
the changes that they undergo as a result of culture contact (Trimble 1989). Berry (1997) defined
three contextual factors that affect how people assimilate: voluntariness, mobility and the
expected permanence of the move. Other variables that have been found to influence the
assimilation process include: the home country's political, social and economic environment;
reason for immigration; prior contact with their new society; the route, duration and level of
danger of the immigration j ourney; the host country' s social, political, economic environment
and immigration policies; social attitudes towards migrants; and demographic factors of the
migrant (Cabassa 2003).
Empirical research indicates that assimilation may not be so neat. An individual may
acculturate without the permission of the larger society by learning the social norms through
interaction with others within the culture or through the media, all while not assimilating. Keefe
(1979) found evidence that migrants became acculturated into American cultural norms while
their social ties were unaffected. Researchers in psychology found that groups could maintain
their traditions while simultaneously assimilating pieces of the dominant culture that fit with
their traditional viewpoints (Trimble 2002). Another criticism of unidimensional assimilation is
that it assumes that people must give up one culture to gain another; it is a zero sum phenomenon
(Cuellar, et al. 1980) .
Recent research on assimilation has moved from unidimensional to multidimensional
models. Berry (2002) argues that there are four categories of acculturation: assimilation,
integration, separation and marginalization. These outcome categories are driven by the degree to
which the migrant maintains his or her heritage culture and the degree to which the migrant
seeks relationships in the new culture (Berry 2002). Thus, if a migrant has a high score on
seeking new relationships and does not attempt to maintain his or her heritage, that migrant is
said to be assimilating. At the other extreme, a migrant who fully maintains his or her heritage
and develops no new relationships results in a state of separation from the new society.
Integration occurs when the migrant simultaneously makes new relationships and maintains the
relationships they had in their homeland. Marginalization occurs when a migrant neither
develops new relationships nor maintains their old ones. Devereux and Loeb (1943), Tyler et al.
(1996), and Teske and Nelson (1974) also argued over the years for a bi-dimensional structure
in the assimilation process.
During most of the twentieth century, migration was studied under the assumption that
the longer migrants live abroad the more they would be assimilated into their new society and the
fewer associations they should have with people in their home country (Guarnizo, et al. 2003).
However in the early 1990's the term 'transnational migrants' was used to describe people who
maintain lives in both their home and host countries (Portes 2001). Hannerz (1998) points out
that transnationalism did not begin in the late twentieth century; it was the study of this behavior
that became prominent. The concept of transnationalism refocused the study of immigrants into
an examination of how they maintain relationships across national borders.
The word 'transnational' has been used for some time in intellectual traditions outside of
the social sciences. The Columbia Journal of Transnational Law was established in 1963 (York)
and Martinelli used the term 'transnational' to describe corporations operating in multiple
countries (Martinelli 1982). The process of transnationalism refers to migrants creating and
maintaining two sets of social relationships, one in their home country and one in their host
country (Basch, et al. 1994; Irzigsohn and Saucedo 2002). Similarly Glick-Schiller et al. define
transnationalism as the "process by which immigrants build social fields that link together their
country of origin and their country of settlement" (Abaza 2001; Glick Schiller 1995 p. 171).
The study of transnational migrants is important for a number of reasons. First, it is
likely that the phenomenon will continue to grow with the development of global capitalism.
Second, previous research on migration has emphasized how migrants integrate into their new
society, however transnational migrants may integrate into society differently. And finally, with
the flow of ideas and remittances, the impact of transnational migrants on both their host and
home countries has been culturally and economically significant (Levitt 2001; Portes 2001).
In anthropology, the concept of transnationalism has three primary features: first, the
social or mental process or behavior under study must cross a political border in some way;
second, the emphasis of the study will be on the distance between two political units; and third,
a multiplicity of meanings must be involved (Hannerz 1998). Portes et al. (1999) identify three
kinds of transnationalism: economic, political and socio-cultural. Economic transnationalism is
exemplified by businessmen and women who have suppliers and markets across national
borders. This category may include migrant laborers who cross national borders on a regular
basis and migrants who return to start businesses in their home country. Political
transnationalism consists of immigrants who participate in activities such as raising funds for
political candidates in their home country while living in the U.S. Socio-cultural
transnationalism consists of displaying the cultural practices of migrants in multiple countries,
such as the world wide export of Ecuadorian folk crafts and music.
Transnationalism has also been divided by the degree of institutionalization that is
involved in the practice. Transnational activities that are created by the state, or large
international organizations are considered to be high on this institutional scale-transnationalism
from above--while smaller, grass root activities are classified as being low on the scale--that is,
transnationalism from below (Portes 1999).
Portes et al. argue that the concept of transnationalism should capture part of the
migration experience that is not part of other phenomena (Portes 1999). It is recognized that
many migrants have occasional contact with their home country. They take trips home for
special occasions, keep in touch with a few close friends they grew up with or send remittances
home (Foner 1997). A number of studies have examined the linkages between previous migrant
groups and their home country (Portes 1999; Schiller and Basch 1995). But in these studies
migrants who maintained ties to their homeland were the exception and 'lacked the elements of
regularity, routine involvement and critical mass'(Portes and Landolt 1999); they were not, as
Glick Schiller et al. (1995) have stated, a part of the metaphor of 'America as a melting pot' and
therefore were discounted from mainstream migrant research. Several authors argue that it is the
degree to which transnational behaviors occur that makes transnationalism a distinct
phenomenon (Basch, et al. 1994; Itzigsohn, et al. 1999; Landolt 2001)."What constitutes truly
original phenomena...are the high intensity of exchanges, the new modes of transacting, and the
multiplication of activities that require cross-border travel and contacts on a sustained basis"
(Portes 1999) .
The definitions of transnationalism offered by Portes and others are based on empirical,
ethnographic research (Basch, et al. 1994; Glick Schiller, et al. 1992). Ethnography is essential
for identifying and documenting social processes and for establishing the content of those
processes. Once a concept, like acculturation or transnationalism, has been documented and its
components identified, we need reliable measurement in order to test theories about what causes
it and what it causes. This step has proved more difficult than might have been anticipated.
One important feature of transnationalism, noted by many researchers, is the variety of
practices it encompasses and the range of participation in these practices (Levitt, et al. 2003).
Itzigsohn and Saucedo measured transnationalism in a large, multi-site survey on public
activities (Irzigsohn and Saucedo 2002). They asked respondents to rate their participation in the
following activities: participation in hometown associations, sending money home for
community proj ects, traveling to public festivities, participation in local sports clubs, and
participation in charity organizations that are linked to their country of origin (Irzigsohn and
Saucedo 2002). A survey in Vancouver, Canada, asked nearly 2000 migrants how and how
frequently they stay in touch with friends and family, frequency of travel to the home country,
whether they owned either a business or property at home and whether they provide Einancial
assistance to family and friends who remained in their home country (Hiebert and Ley 2003).
Cohen et al. conducted a survey in Mexico that measured the ties people had to previous
migrants, how migrants paid for their trip, where they lived once they arrived in the U.S., the
number of years the migrant had remitted funds and the number of migrants the household had
previously sent (Cohen, et al. 2003). Political transnationalism has been measured with
questions about a migrant's membership in political parties or charity organizations at home,
donations to political parties or to community proj ects, and participation in elections or civic
associations at home (Guarnizo, et al. 2003).
Depending on the definitions of transnationalism the degree of its prevalence varies
considerably. Itzigsohn and Saucedo's (2002) survey of three communities (Dominicans,
Salvadorans, and Columbians) found that 22% of migrants participated in a transnational activity
on a regular basis, with 74% of the migrants sending money home regularly. Hieber and Ley' s
(2003) study found that: 41% keep in touch with their friends in family in their home county at
least monthly; 20% travel to their home more than once a year; and 25% either own property or a
business in their home country. Cohen et al. (2003) found that 25% of respondents scored high
on their composite measure of transnational behavior.
While these numbers estimate the degree to which migrants participate in transnational
activities they also indicate the variability in migrant behavior. The concept of broad versus
narrow transnationalism captures this idea (Hiebert and Ley 2003; Itzigsohn, et al. 1999).
Narrow transnationalism is defined as the continuous participation in transnational activities by
migrants. Broad transnationalism is the occasional participation in these activities.
Macro-level theories of transnationalism have related the phenomenon to globalization of the
economy and of political forms, improvements in communication and transportation
technologies and changes in international laws (Basch, et al. 1994; Faist 2000; Portes 1999;
Portes and Landolt 1999; Smith and Guarnizo 1998). Globalization of the economy has been
going on since capitalism began replacing feudalism, but the pace accelerated with the collapse
of the Soviet Union. Global political changes include decolonization, the universalization of
human rights, and nation building proj ects that focus on developing loyalty to both home and
host countries (Schiller and Basch 1995; Smith and Guarnizo 1998). These factors are combined
with deteriorating economic and social standards of living created when companies move
factories into lesser developed countries without adding infrastructure (Basch, et al. 1994; Faist
2000). Racism in both the U.S. and Europe is also cited as a contributing factor as it prevents
migrants from integrating into the destination society (Basch, et al. 1994). These factors do not
motivate transnational migration on the personal level but they provide the preconditions for
transnationalism that were not present to previous waves of migrants.
Micro-level theories of transnationalism focus on explaining the motivation for
transnational migration (Irzigsohn and Saucedo 2002). Linear theories argue that migrants
attempt to maintain ties with their home county. Resource dependence theories assert that
migrants can not behave transnationally until they have the resources to do so, i.e. there is a base
level of income that is needed to maintain these international ties. Reactive theories make the
case that the negative experiences of moving to a new society influence migrants to maintain ties
to their old society. The variables that have been used to test these theories are described below.
Note that in each of these studies transnationalism has been measured differently. Nevertheless
there are some interesting similarities.
In all studies reviewed, transnationals are typically young married men who have
children in their home country (Guarnizo, et al. 2003; Irzigsohn and Saucedo 2002). Two studies
support the resource dependence theory of transnationalism and found that a basic level of
income was required by the migrant to support transnational activities (Guarnizo, et al. 2003;
Hiebert and Ley 2003). Employment status (having vs. not having a job) has also been identified
as significant, though not the kind of job held (Irzigsohn and Saucedo 2002). The goods owned
by a family appear to be related to transnational activities (Cohen, et al. 2003), though this
variable may be a proxy for income. The number of years in the U. S. is positively and
significantly related to transnational practices (Guamizo, et al. 2003; Hiebert and Ley 2003). In
Canada (Hiebert and Ley 2003), the interview language, if not English, is associated significantly
with transnational practices indicating that the migrants felt more comfortable using their native
language. On the other hand, citizenship status of the migrant was not significant (Guarnizo, et
al. 2003; Irzigsohn and Saucedo 2002). The period of time that an immigrant has been in the
country was found to be very significantly related to degree of transnational behaviors exhibited
(Hiebert and Ley 2003). Travel to their home country and receiving visitors from their home
country were more frequent if the migrant had been in the host country more than 10 years. All
transnational activities were more frequent if the migrant had been in the host country for less
than 10 years (Guamizo, et al. 2003). The expectation that the migrant would someday return to
their home country was also found to be a significant predictor of transnationalism (Guarnizo, et
al. 2003; Irzigsohn and Saucedo 2002; Landolt 2001). Only one study examined the idea that
transnationalism may be a reaction to the stresses of the assimilation process. Irzigsohn and
Saucedo (2002) found that an increase in experiences of discrimination led to increased
transnational behavior and that transnationalism was positively related to an increasingly
negative view of the host society.
Education was especially significant in this research, with a positive relationship to
transnational behavior (Guarnizo, et al. 2003; Irzigsohn and Saucedo 2002). Education is also
related to occupational status, and thus the resources available to the migrant. However, there
can be incongruity between education and occupational status, when migrants who are well
educated are overqualified for their current work. Itzigsohn and Saucedo (2002) therefore argue
that education may be picking up other resources, available to the migrant in the form of social
networks or wealth that are not captured in occupation. Another point of view is that the more
educated a person is the easier it is for them to establish new ties in their host country. In any
event, education facilitates the migrant' s integration into the host society (Borj as 1987).
Research in transnationalism examines the motivations for participating in transnational
activities or behaviors. Unfortunately it has not measured how the culture of migrants changes
during this process. Acculturation theory examines the similarity between the culture migrants
and that of the society into which migrants move. However, acculturation theory is based a
model of migration where migrants move from their home country to their host country and
attempt to set up a new life. They may return infrequently, but as a whole their lives are
conducted in the country of origin. There is some evidence that assimilation and participation in
transnational activities are not mutually exclusive (Basch, et al. 1994; Irzigsohn and Saucedo
2002). Itzigsohn and Saucedo (2002) found that there is a marginally significant but positive
relationship between the time spent in the US and a migrant' s participation in transnational
activities. Transnationalism does not subtract from assimilation, nor assimilation from
transnationalism; they are simply additional layers of behavior seen in modern migrants (Basch,
et al. 1994). The research reported here examines changes in the culture of migrants who move
back and forth between Mexico and the US to such a degree that they are actively involved in
both cultures. To understand these changes I compare the variation, structure and composition of
three cultural domains in three populations. I hypothesize:
* H1: The cultural models of migrants are a mix of elements from the home and host
society's cultural models.
o Alternative 1: The cultural models of transnational migrants will have elements
that are neither in the home nor host country's cultural models.
* H2: The structure of the migrants' cultural models will be more fragmented than those of
people in either the home or host culture.
Besides providing insight into post migration changes of transnational migrants, this
research is also about cultural domains. The idea of cultural domains was developed by cognitive
anthropologists and psychologists as a way to understand the relationship between culture and
thought. Cultural domains are a member of the larger class of cognitive schemas but are defined
as schemas that are shared by large segments of a culture and created with systematically
collected data, such as pilesorts, freelists or triad tests. These cultural domains have proven
useful in understanding how people structure plants, animals, foods, emotions, and material
goods, among others. However most of the research using cultural domains has been with
populations that are strongly situated in one culture. With the interconnectedness of today' s
world, people migrate repeatedly between cultures to such a degree that they are influenced by
multiple cultures. This research examines how these domains are affected when people have
multiple cultural influences. This chapter presents a short history of the development of cognitive
anthropology and specifically the idea of domains and argues for their usefulness in
understanding the changes in migrant culture.
An early attempt to define cognitive anthropology was made by Tyler (1969: 3) stating
that; "Cognitive anthropology seeks to answer two questions: What material phenomena are
significant for the people of some culture; and, how do they organize these phenomena?" This
definition developed from the crisis in ethnographic authority that began with the Redfield-Lewis
debate in the 1940s and 50s (see below) and continued with Freeman' s critique (Freeman 1983)
of Mead' s (1928) work in Samoa and Weiner' s (1983) re-evaluation of Malinowski's work in
the Trobriands (Malinowski 1922). In all of this, ethnographers were finding behavior and
beliefs quite different from those reported in the work of previous anthropologists. The Redfield-
Lewis debate gave the discipline incentive to find reliable methods for recording cultural data.
By the time the crisis reached the 1980s, however, many of the computationally intensive
methods that cognitive anthropologists had developed were falling into disfavor. More recently,
with the development of easy-to-use software, these methods are making a comeback. The
method of cultural schema analysis, however, does not require prodigious computer skills and
has grown in influence over the years (Bernard 2007)
Recent definitions move beyond the basic organization of cultural knowledge and
emphasize the idea of cultural models. The development of schema theory and cultural models
shifted the problem of modern cognitive anthropology to, "how, and from what manner of
evidence, to reconstruct the cultural models people use but do not often reflect on or explicitly
articulate" (Quinn and Holland 1987:14). Perhaps the broadest definition of the discipline is
given by D'Andrade (1 995: 1): Cognitive anthropology is the study of the relation between
human society and human thought. The cognitive anthropologist studies how groups conceive of
and think about the obj ects and events which make up their world..." The rest of this chapter
traces the development of cognitive anthropology as the basis of cultural schema analysis.
The stimulation for the creation of cognitive anthropology came from a separation of the
material evidence of culture and the knowledge that is needed to create this evidence. This split
was found in Goodenough' s (1957 P. 167) definition of culture, stating that, "A society's culture
consi sts of whatever it is one has to know or believe in order to operate in a manner acceptable to
its members." This definition diverged from earlier definitions of culture that emphasized the
customs, artifacts and oral traditions of a society and instead focuses on the knowledge that
people must have to create and understand these cultural artifacts (Quinn and Holland 1987).
The cognitive revolution was underway in other social and psychological sciences at the
time when cognitive anthropology was taking root. This revolution heavily influenced the new
sub-discipline with a change from the traditional perspective of strictly studying observable
behavior to one that explored the internal workings of the mind. Cognitive anthropologists saw
culture as the internal representations of experience that was the mediating factor between
sensory data and the innate categories of rationalism (Foley 1997). Culture moderated how
people assigned experiences and how they classified the information from their senses. This idea
was equivalent to Chomsky's grammar for language. In fact, in his earliest research, Franz Boas
became interested in the relationship between the environment and how it is perceived.
Specifically he studied how Eskimo perceived the color of ice and water (Boas 1887) and
concluded that culture creates a bias in the perceptions of people.
The New Ethnography
The cognitive revolution laid the foundation for cognitive anthropology. However, when
the reliability of ethnography came into question early in the 20th century, the discipline was
called upon to provide validity for ethnographic studies. This problem was the result of
increasing numbers of disparate accounts from ethnographers researching the same people. The
most famous of these differences became know as the Redfield-Lewis debate. Robert Redfield
wrote a monograph from work in the Mexican town of Tepoztlan in the 1920s (Redfield 1930).
A team of ethnographers led by Oscar Lewis returned to Tepoztlan in the 1940s and found many
points of difference from the reports by Redfield. The differences seemed to be more than could
be accounted for simply by the lapse in time, which raised the problem of reliability in cultural
anthropology (Colby 1996). Anthropology was plagued by the question of how anthropologists
could give an accurate account of a culture when the biases of the ethnographer interfered with
their ability to create accurate records.
To tackle this problem, methods and ideas were borrowed from formal linguistic
analysis--methods that would, it was hoped, eliminate the bias created by the ethnographer' s
own cultural categories in the conduct of ethnography- (D'Andrade 1995). The focus of these
methods became a culture's terminological system, its language. Analyzing a language would
not reveal the entire cognitive world of a people; however most significant features of a culture
would have to be communicable and therefore encoded within the language (Frake 1962).
Methods developed from the Prague Linguistic Circle emphasized the study of the
structure of language. Phonology used this view to investigate the factors that made one chunk of
sound different from other chunks of sound. They argued that the units of sound could only be
identified in terms of their relationships with other units of sound. Componential or semantic
feature analysis carried over into anthropology with the study of cultural domains in the same
manner. They attempted to describe the features of a category that differentiate the terms within
it (Goodenough 1956). Features or components were the "dimensions of meaning underlying
the general domain" (Tyler 1969:8).
Lead by Lounsbury and Goodenough early efforts were labeled 'ethnoscience' or 'the
new ethnography' and attempted to capture the cultural construction of reality by examining how
people labeled different parts of their world (Goodenough 1956; Quinn and Holland 1987).
Researchers examined the structure of semantic domains, which were "a class of obj ects all of
which share at least one feature in common which differentiates them from other semantic
domains." (Tyler 1969:8) An example of a semantic domain is furniture and would consist of
items like chairs, tables, sofas, beds, and dressers.
Wallace and Atkins list five steps in a componential analysis of the domain of kinship
terminology (Wallace and Atkins 1960). (1) Collect a complete set of terms, (2) Record the
terms (in this case, in kin-type notation). (3) Identify two or more conceptual dimensions. (4)
Define each term as a set of components. And (5) write a statement of the relationship among the
terms of the structure of the system. While this method was specific to the analysis of kinship
terminology it can be (and was) adapted to the study of other domains.
A simple example of a classic feature model is given by Tyler (1969) in an examination
of the taxonomy of American English livestock terms. First the terminology is gathered: cattle,
cow, bull, horse, mare, stallion, filly, sheep, ewe, ram, lamb, swine, sow and boar, and so on. The
defining features are identified as the type of animal (horse, pig, cow or sheep), gender (male,
female, neuter), and maturity (immature, mature, adolescent, baby). A notations system is used
to define each term, such as H $ M ~1(horse, male, adult). The researcher then attempts to define
each term using these components. If successful, the native terms are decoded and laid out in a
paradigm. This use of the term "paradigm," not to be confused with Kuhn' s, involved "any set of
linguistic forms wherein (a) the meaning of every form has a feature in common with meaning of
all other forms of the set, and (b) the meaning of every form differs from that of every other form
of the set by one or more additional features"(Lounsbury 1964b,:193).
There were other types of semantic arrangements used in ethnoscience. Taxonomies
differed from paradigms by creating levels of contrast, such that a horse, for example, is a kind
of livestock. Another arrangement was the tree, where, "features in a tree are ordered by
sequential contrast of only one feature at a time" (Tyler 1969: 10). A number of researchers used
these techniques to understand complicated systems of terminology. Goodenough (1965)
examined American English kinship terms while Lounsbury (1964b) discovered that the kin
terminology of the Seneca Indians in western New York state had four distinguishing features:
polarity, sex, bifurcation and generation distance.
Componential analysis was used with considerable success in the analysis of kinship
terminology. These analyses provided working models that could be used to predict when a
person would label someone with a particular term. The idea of finding the set of rules which
would accurately decide when a term could be applied to an obj ect is a marked advance in the
history of anthropology. However, Burling (1964) pointed out that componential analysis
actually has two goals. Besides understanding the rules that would lead to the correct application
of a term the method also sought to provide an understanding of the criteria used by native
speakers to decide when to use a particular term. Burling, as well as others, found that there may
be multiple solutions that satisfy the first goal and that finding a working solution may not reflect
the rules applied by a native (Burling 1964; Quinn and Holland 1987) Componential analysis is a
useful start for understanding cultural domains, but the understanding is insufficient for
understanding the native point of view (Quinn and Holland 1987).
Semantic Networks and Ethnobotany
Even with the possibility that feature analysis might not be sufficient, the idea was
extended into domains besides kinship. Metzger and Williams (1966) examined the features of
firewood terminology among the Tzeltal in Mexico and found that the methods used in the study
of kinship terms were lacking. Their solution was the frame-and-slot method of questioning.
Informants were asked, systematically, to state whether "_ is a type of tree. Some answers
would make this sentence true and others wouldn't. By understanding the possible sentences that
a term could be used in, one could understand the term the way that natives would understand it
(Metzger and Williams 1966). Other domains were studied using similar methods (see Black and
Metzger 1965; Frake 1964) but most importantly, these techniques could be used to
systematically and reliably obtain knowledge about any domain (D'Andrade 1995: 61).
Semantic network analysis built on these methods by not only examining the situations that
a word could be correctly used in, but also explored how categories of words were related (Frake
1964; Werner 1987). In semantic network analysis the responses to questions were interlinked
through various queries. These connections produced networks of linked propositions that could
describe everything in a culture, at least theoretically.
At the same time as the feature model was being developed, Conklin (1954) began his
work in ethnobotany. This sub-discipline focused on native taxonomies of plants and animals.
He found that these taxonomies usually have no more than five levels (Berlin, et al. 1973).
D'Andrade (1995) argues that this is a function of the restrictions on human memory more than a
reflection of the nature of plants and animals. Similar limitations on size were found when the
taxonomic model was applied to other domains (Brown, et al. 1976)
The next development in cognitive anthropology draws from work by the philosophers
Wittgenstein and Austin (Lakoff 1987). They argued that words and categories do not have a
single set of features that are common to all the pieces normally associated with a category. They
also hypothesized that some members of a category were better examples of the category than
other members. In other words, some were more central to the meaning of the category.
Loundsbury (1964a) found evidence of these aspects in his work on kinship terminology. He
noted that the meaning of kinship terms were based on the genealogically closest relative but the
meaning was extended for all of relatives in that class. These types of categories later were
known as generative categories because the category can be generated by applying rules to the
focal member. Berlin, Breedlove and Raven (1974) formulated a similar idea in their
investigation of taxonomic systems. They found that terms have a basic and an extended range.
The basic range includes those items that form the core of a class. The extended range are those
items that are not core items but fall into one class more than they do any other class.
In another line of research, Berlin and Kay (1969) set out to disprove the theory of
linguistic relativity advocated by Sapir and Whorf-that language influenced the way people
perceived the world and that language did this in an arbitrary fashion. In the strong form of
relativity, every language divided the universe in different ways and had no obj ective base.
We dissect nature along lines laid down by our native languages. The categories and
types that we isolate from the world of phenomena we do not find there because they
stare every observer in the face; on the contrary, the world is presented in a kaleidoscopic
flux of impressions which has to be organized by our minds--and this means largely by
the linguistic systems in our minds. (Whorf quoted in Carroll 1956: 213)
Berlin and Kay (1969), found that the way the color spectrum is divided into linguistic terms is
not arbitrary. It turns out there are eleven basic colors that can be distinguished by people across
cultures and languages, even if a language does not have color terms to make the differentiation.
Additionally they found that languages added these categories in a partially fixed order. Similar
to previous researchers' findings these color categories had focal colors or prototypes that were
the best example of a particular color. All of these ideas coalesced into prototype theory (Berlin
and Kay 1969; Rosch 1975; Rosch 1977; Rosch 1978a; Rosch 1978b; Rosch and Mervis 1975).
Casson (1983) defined a prototype as, "a stereotypic, or generic, representation of a
concept that serves as a standard for evaluating the goodness-of-fit between schema variables
and elements in the environment." (P. 434) Rosch (1978b) brought together these ideas and
hypothesized that the basic level terms of Berlin and Kay correspond to basic level psychological
obj ects. These obj ects are perceived through the prototype, the stereotypic whole that makes up
the obj ect. A number of experiments confirmed the idea that prototypes had a psychological
reality (see Rosch, et al. 1976).
Prototype theory helped move cognitive anthropology past basic feature analysis to study
more complex systems of features. The next development expanded the idea of prototypes into
schema or schemata. While Kant used the term schemata, its first use in its modern sense was by
Bartlett (1932) when he stated that schema are used to give 'an impression of the whole'
(Bartlett 1932 quoted in Casson 1983: 430). Bartlett developed the idea of schemata from his
work on memory. Using two different methods he discovered that these impressions or schemata
influence both the comprehension and the memory of an event in culturally specific ways
(Bartlett 1932). Similar ideas were developed by Schank and Abelson with their work on how
language is understood (Schank and Abelson 1977). They developed the idea of scripts as a
series of typical actions in an event (Schank and Abelson 1977). A small number of scripts were
used to understand many different situations. Memory of an event consists of the typical script
for that event, plus markers for any deviations from that script. Besides schemata and scripts
researchers have used other names to refer to this concept including: 'frames', 'scenes',
'scenarios', 'gestalts', 'active structural networks' and 'memory organization packets' (Casson
More recent work on schemas define them as "conceptual abstractions that mediate
between stimuli received by the sense organs and behavioral responses (Casson 1983: 430)."
These were mental entities that are used to classify obj ects and understand relationships and
events. Quinn (1996) points out that schemas may also be used as a tool to reason through
typical problems found within a culture. Schema theory moved the study of cultural knowledge
away from the static forms of ethnoscience and gave these knowledge structures a dynamic
nature, capable of being used to process information (Rice 1980). They can be used in perception
to recognize an obj ect or event as well as in memory by organizing the event along a set of
A number of authors have investigated how people construct new schemas (Collins and
Gentner 1987; Lakoff and Johnson 1980; Rumelhart and Norman 1981). Collins and Gentner
(1987) examined how people used multiple domains and analogies when one established domain
does not cover the entire range of phenomena in a model of evaporation. The rules for the
transition from one state to another were transferred to the new domain to explain some subset of
the phenomena. Other analogies were used for different subsets. They also noted that people
coordinated the various analogies to various degrees of consistency. Analyzing natural discourse,
Price (1987) found that a number of cultural models were combined in explanations of the origin
of illness in Ecuador. In an investigation about the use of text and schemas, Quinn (1996) found
that words contained the cultural knowledge of a society, but also allowed variation in how this
knowledge was used.
Schemas have a number of interesting features. One aspect of schemas is that they leave
some information out (Rumelhart and Ortony 1977). This information can be filled in by the
specific situation or if no information is available then a default value is used (Fillmore 1977;
Minsky 1975). Ricc (1980) demonstrates this aspect of schemas by showing how people modify
unfamiliar stories to take on the form of a more typical, known schema (this was also, in fact,
Bartlett' s key finding (Bartlett 1932)). Schemata form networks with other schemata that can be
used to understand changing scenes (Rumelhart and Ortony 1977). These types of schemata have
been labeled scripts by Abelson (1976). Schemas also have various levels of abstraction
(Rumelhart and Ortony 1977). Some schemas simply help us perceive and classify simple
shapes and colors. Other types of schemata aid us in understanding complex events (Casson
1983). The more abstract schemas have been termed foundational, while particular instantiations
of schemas are called models (Lakoff 1987; Shore 1996). The schema concept has a hierarchical
element such that larger concepts or higher level schemas are composed of multiple lower level
schemes (D'Andrade 1995). The lower levels are less fixed and have variables that can accept
different instantiation as dictated by the environment (Mindsky 1975). The variables in the
lower level schemas are restricted in the type of input they take by the rules created by the higher
level schemata (Rumelhart 1980).
Schemas exist at different levels of sharing (Shore 1996). There may be universal
schemas that are shared by everyone, everywhere. There are certainly idiosyncratic schemas that
are restricted to a single person (Rice 1980). Between these two extremes are schemas that are
shared within a culture. "A cultural model is a cognitive schema that is inter-subj ectively shared
by a social group... A schema is inter-subj ectively shared when everybody in the group knows
the schema, and everybody knows that everyone else knows the schema..." (D'Andrade 1987: 112
- 113) One of the consequences of the sharing involved in cultural models is that large amounts
of information can be left out of communications between people if they share the same cultural
models (D'Andrade 1995).
Shore (1996) notes that the sharing of cultural models also means that these models are
constrained by the social environment. People may reinforce these models both positively and
negatively. Shore (1996) makes the distinction between personal and conventional models with
the distinguishing feature being the degree to which they are shared. One consequence of this
division is that conventional models can be used by people to form personal mental models. This
allows some variability in the degree to which people know a particular cultural model. If they
learned the model entirely, their personal and cultural models will match. If they learned the
model only partially or have modified their personal model with other experiences then the two
models may not match up (Shore 1996). Both D'Andrade (1995) and Shore (1996) note that
these cultural models are strictly a mental entity. They do not exist in the world. However they
may influence everything from expectations about family life to the operation of public
Shore (1996) has laid out two different classification structures for cultural models. In
one structure he divides the models into two large categories based upon their structure -
linguistic and non-linguistic. Linguistic models are concerned with aspects of verbal
communication and include scripts (Garvey 1977), propositional models (Langer 1957), sound
symbolic models, lexical models (D'Andrade 1990), and grammatical models (Lakoff 1987;
Talmy 1983). The term "nonlinguistic models" is a catch-all for other types of models, including:
image schemas (Johnson 1987; Lakoff 1987), olfactory models (Sperber 1975), sound image
models (Basso 1985) and visual image models (Richard 1974).
The second system of classification found in Shore's work is based on the cultural
model's function. This system had three major divisions: Orientational models,
Expressive/conceptual models and task models. Orientational models provide a common
framework for people to orient themselves to the physical environment, time frames and the
social environment (Evans-Prichard 1940; Morphy 1991). Also included in this category are
models that aid in diagnosing important phenomena (Lakoff 1987). Expressive / conceptual
models "crystallize for communities important but otherwise unspoken understandings and
experiences." (Shore 1996:64) This category includes classification models like the taxonomic
models from early ethnoscience, exemplars, ludic models (games, sports and humor), rituals and
folk theories (Kempton (1987) provides an excellent example of a folk theory for thermostat
usage). The final category includes models for accomplishing different tasks. These include
scripts, recipes, checklists, mnemonic models and persuasion models (Schank 1991).
While the idea of cultural models was a significant advancement for the discipline a
systematic manner to collect these models was still needed. This development was provided by
the cultural consensus model developed by Romney, Weller and Batchelder (Romney, et al.
1986). The basic idea is that when anthropologists speak with informants they should assume
that the informants' statement have a probability of being correct. The more informants who
make the same statement, the more likely it is to represent their culture. Boster (1986) found
evidence of this in his work with the naming of different types of manioc among the Aguaruna
Jivaro. He was able to identity informants with the most knowledge by their agreement with
The Romney et al. (1986) model is based on the following assumptions: 1) all informants
come from a common culture, 2) each informant' s answer is given independently from others so
that any correlation between the answers is a result of their knowledge of the correct cultural
answers, and 3) each informant has a fixed degree of knowledge for all of the questions
(Romney, et al. 1986). The method has been tested for true/false, multiple choice and fill-in-the-
blank questions. The results of the model include an estimate of the degree of cultural
competence for each respondent and a culturally correct answer key--that is, the answers from
an informant who best knows a particular cultural domain, given the set of questions that define
The fact that cultural knowledge is shared means that, if the model's assumptions are
met, very few informants are needed when cultural consensus is high about a domain in order to
assess cultural knowledge. Handwerker and Wozniak (1997) argue that because cultural data are
socially constructed, sampling strategies for these types of data should focus on selecting
informants who have knowledge within the domain in question. The minimum number of
participants needed to describe a cultural domain is a function of their cultural competence, the
minimum acceptable confidence level needed in a particular study and the proportion of
questions asked that will need to be decisively classified (Romney, et al. 1986). The larger the
proportion of questions asked that need to be classified and the higher the acceptable confidence
level, the larger the number of informants needed. Even in the most extreme case (asking for a
.999 confidence level and .99 proportion of questions classified) only 23 informants are
necessary if the average level of cultural competence is .6. At an average competence level of .7
only 16 informants are needed (Romney, et al. 1986).
It should be noted that there are several differences between cultural models, as
developed by Strauss and Quinn (1997), and cultural domains (Romney, et al. 1986). The first
difference is methodological. Cultural models are usually developed with extensive textual data,
while cultural domains are created from systematically collected data such as pilesorts or triad
tests (Garro 2000). A more fundamental difference is each theory's view of cultural knowledge.
Boster (1987) noted that cultural consensus analysis assumes a particle theory of cultural
knowledge. Cultural knowledge is made up of discrete pieces that have certain properties.
Cultural models developed out of schema theory, which focused on the interrelationships of
knowledge and how these networks influenced the acquisition and storage of knowledge (Strauss
and Quinn 1997: 6).
Cultural Models and Behavior
The ideas reviewed above come from the examination of the relationship between culture
and how it is reconstructed inside the mind. Another line of research examines how these mental
constructions are related to behavior. Holland (1992) argues for three ways that culture may be
related to behavior. The first is that culture is simply the external coating of deep-seated human
needs. This view places the motivational power for behavior in "more fundamental
psychodynamic, materialist, or even structural substrate" (Holland 1992:63). The second
relationship between culture and actions gives culture primary motivational force to influence
human needs. The third connection between culture and action derives from the neo-Vygotskian
development approach. In this approach the person internalizes the cultural models which in turn
become part of his or her mental functioning. The motivation is created by the process of
learning the cultural symbols and meanings of a particular domain. Holland found that in the
realm of romance the more expertise a particular person had the more the person identified
themselves with romance.
D'Andrade links cultural schemas to action through motivation (D'Andrade 1992). He
argues that schemas have certain properties that make them capable of providing motivation.
First they are processes. These processes create interpretations of the world that have different
degrees of schematicity. The more schematicity an interpretation has the more likely it seems
typical. Schemas can also act as goals (Quinn 1996). However, while they can provide
motivation, they do so in differing degrees. Schemas are also hierarchical, such that the
interpretation from one schema can be passed on to other schemas. The highest level schemas
provide the more general goal setting abilities. Mid-level motives rarely provide independent
motivation and usually function in the context of higher level schemas. The lowest level schemas
never provide motivation except in the context of higher order schemas.
Quinn (1992; Quinn 1996) builds on D'Andrade' s model, arguing that some schema's
become the highest level schemas by supplying people with an "understanding of ourselves"
(Quinn 1992: 91). Schemas are not created with this ability to give this understanding, but
acquire it through critical experiences, often in childhood and adolescence. Strauss (1992)
agrees with D'Andrade in that schemas often act as goal setting devices. However she argues
that high level cultural models do not necessarily provide the motivation for behavior. She found
that different ways of believing in the model was the key force driving certain behavior (Strauss
Cultural consonance builds on cultural consensus analysis and provides a method for
measuring the relationship between the cultural model and actual behavior. Cultural consensus
analysis produces a cultural model if there is sufficient agreement between the informants.
Cultural consonance measures the degree that behavior is related to the ideal model (Dressler
1996; Dressler, et al. 1997; Dressler and Bindon 2000). Cultural consonance has been used to
find relationships between social support, blood pressure and perceived stress (Dressler, et al.
A Critique of Cognitive Anthropology
In 1968 Marvin Harris (1968) wrote an insightful critique of cognitive anthropology as it
was practiced at the time. Some of the issues he pointed out have been addressed with the
development of the discipline; others have not. Harris noted six problems with cognitive
anthropology. The first four were products of a newly developed sub-discipline and have been or
can be overcome with later work.
Harris addressed his critique to what he saw as a conflation of the emic and the etic
components of culture. Derived from Pike's (Pike 1954) analogy with phonemics and phonetics
the emic-etic concept divides anthropological knowledge into statements that are significant
within the culture under study Knowledge of this sort is based in the "logico-empirical systems
whose phenomenal distinctions or "things" are built up out of contrasts and discrimination
significant, meaningful, real, accurate, or in some fashion regarded as appropriate by the actors
themselves" (Harris 1968: 571) The other type of knowledge, etic, does not depend on the
understanding of the native actors. They are appropriate to the scientific community. They
"cannot be falsified if they do not conform to the actor' s notion of what is significant, real,
meaningful, or appropriate....(and are) verified when independent observers using similar
operation agree that a given event has occurred." (Harris 1968: 575)
The first critique is that emic studies do not further the scientific purpose of
understanding sociological differences and similarities. However, Berlin and Kay' s (1969) work
with color terms provides us not only with knowledge about the distribution of color terms; it
also shows how the evolution of complexity in color term systems mirrors the evolution of social
complexity. This research used emic color terms advance research in cognitive anthropology.
The second critique is that emic statements are often interspersed with etic statements in the
description of economic processes. While this was true in early ethnographies, modern work is
very strict about the types of data used in cognitive anthropology. The third critique is that emic
analyses assume that models of ideal behavior are superior to models of actual behavior. Work
by D'Andrade (1992) Dressler (1996; Dressler, et al. 1997) Holland (1987) and Quinn (1996)
have rigorously examined the relationship between behavior and mental models. The fourth
critique was that cognitive studies dealt only with trivial phenomena. In my view, this was never
true, though apparently, it was easy in the 1960s to give short shrift to studies of firewood and
color terms. In any event, it is certainly not true now. This paper has cited many studies that
have expanded the body of literature in cognitive studies.
Harris's most important critique--that culture is more than mental codes--remains
correct. Harris argued that studying cultural knowledge in isolation from the structural and
infrastructural environment makes it difficult, at best, to develop explanations for differences
across cultures. Additionally, if culture is narrowed to knowledge then the study of past p eoples
is severely limited by the inability to speak with them. Harris also argued that ambiguity in
culture is made insignificant by the methodology of cognitive anthropology. Methods in the
discipline study the norms that are developed from many informants. This excludes the issue that
no one informant may have a complete set of norms or even a standard of behavior that similar to
the cultural norm. This critique was also obviated by the work, after 1975, on intracultural
variation. However cognitive anthropology brings a methodological rigor to the study of that part
of culture that can be verbalized.
Previous measures of cultural change rely heavily on language usage patterns and
associations. The instrument used in this research to measure acculturation, the ARSMA II, has
thirty questions, of which half are about a person' s language. These instruments provide a useful
way to collect data from large samples because of their brevity. However this comes at a cost to
their ability to show the patterns of culture change. The power of constructing cultural domains
is in their ability to provide systematic and detailed insight into the elements and structure of
Cultural consensus analysis tells us that people have different degrees of expertise within
a domain. This may be because people have a particular interest in the topic or they may know
more about a domain because of their experiences in life. To understand how this cultural
knowledge is affected by their repeated migration patterns I will examine the relation between
the knowledge of a domain and other measures of their experience. Two previously developed
scales measure some very important parts of the migrants experience, namely how much they
have assimilated and what their degree of transnationalism is. The Acculturation Rating Scale for
Mexican Americans (ARSMA-II) (Cuellar, et al. 1995) measures their degree of assimilation and
Cohen' s transnationalism score (Cohen, et al. 2003) measures their degree of transnationalism.
These two measures come from very different theoretical bases, as discussed in the previous
chapter. Relating the knowledge that migrant have to these measures may indicate the validity
that these instruments have in measuring some aspect of migrant culture. Additional factors have
been found to be significant in a migrant' s experience (see Massey and Espinosa 1997). While it
is not possible to measure all of these, basic demographic information was collected, including:
age, gender, marital status, number of children, job, and the number of years of education
respondents had received.
* H3: As the degree of acculturation (as measured independently) increases among migrants,
domain knowledge will also increase.
o Alternative H3: As the degree of transnationalism increases among migrants,
domain knowledge will increase.
* H5: Age will increase knowledge across domains.
The design of the research required data from three populations migrants from Mexico
living in the U.S., Mexicans living in Mexico and Americans living in the U.S. The obj ect was
to examine the influence of U. S. and Mexican culture on some aspects of migrant culture. The
first section of this chapter describes the research design. The second section describes the three
locations where I collected data and why these locations were selected. The sampling strategy is
also explained. The third section outlines the methods used to collect the data. To understand the
culture of migrants and how it is similar or different from that of nonmigrants requires a method
that provides a precise and detailed snapshot of culture. The method would have to be systematic
so that it can be repeated on various groups. Since I would not be able to interview all
participants myself, the method would have to be easy enough to train others to administer. The
data required for assessing cultural domains meet these criteria.
There are two important aspects of my research design. First, I am collecting data on three
populations: transnational migrants from Mexico living in the U.S (Migrants), non-migrant
Mexicans living in Mexico (Mexicans), and Anglo Americans living in the U. S (Americans).
Second, I am collecting data on not one cultural domain but three. This section will explain why
these choices were made.
There is evidence for the effect of the home culture on consumer decisions of migrants
(Wallendorf and Reilly 1983) but this is clearly only part of the story about how the consumer
choices of migrants change. A fuller picture of how the cultural domains of migrants change with
regard to consumer choices involves understanding the domains in both the home and host
countries. Without examining these two other populations we cannot tell the influences from
their original culture or their destination culture and we cannot discern what may be unique to
the migrant lifestyle. Because I focused on migrants from Mexico, cultural domains are
identified for them as well as for Mexicans in Mexico and Americans living in the US.
Three cultural domains were chosen because of the possibility that one domain may not
show significant reaction to the influences of multiple cultures. Work on cognitive domains in
psychology indicates that cognitive abilities may be domain specific (Hirschfeld and Gelman
1994). In other words, domains are not simply different realizations of a single structure that
operate on the same set of principles. Instead, each domain has a unique set of rules by which it
operates. There is some evidence that some cognitive domains are more likely to be transmitted
between people (Boyer 1994). While there is no support for the idea that the idea of domain
specificity would translate directly to cultural domains I will assume for now that not all domains
are equal. I will take this issue up later in the analysis. Therefore, to play it safe and not rely on a
single domain, three cultural domains were identified for each population. The three domains are
(1) items owned when a person has led a good life (material goods), (2) typical foods, and (3)
typical pastimes. These domains have proved useful in Dressler's work in Brazil (Dressler, et al.
1997; Dressler and Bindon 2000) on cultural consonance.
Data on the three cultural domains were collected from each of the three populations
using freelists, unconstrained and constrained pilesorts, and ethnographic interviews. Basic
demographic data were also collected on all participants and additional data were collected from
migrants to measure their experiences as migrants. The instruments include a measure of
acculturation and a measure of transnationalism. Both are discussed in detail below. An outline
of the research design is shown Figure 4-1
I conducted pre-dissertation fieldwork during the summer of 2005 in Goodland, Kansas. I
spoke with many migrants there and learned that most migrants in the northwest Kansas-
northeast Colorado region had come from rural areas (pueblos) surrounding the cities of
Chihuahua and Guadalaj ara in Mexico. I selected San Isidro Mazatepec, a pueblo approximately
40 miles south-west of Guadalaj ara, to collect data on Mexicans. This decision was made on the
convenience of arranging housing in Guadalaj ara and on the number of Spanish language
schools in the city, as I planned to take courses during my time in the field. I spent January
through March living in Guadalaj ara but making regular trips to San Isidro (as it is commonly
called). From the end of March to July I lived in Goodland, Kansas collecting data from Anglo
Americans as well as from the migrants there and in the nearby towns of Burlington, Colorado,
and Colby, Kansas.
Examining consumer assimilation processes, Wallendorf and Reilly (1983) noted that
many studies of Mexican and American consumption failed to control for demographic
differences such as age, income, or education level. These intervening variables made if very
difficult to understand if the migrants' consumption behavior was influenced by their home
culture, their host culture or any of the uncontrolled demographic variables. Two of the cultural
domains used in the research, typical foods and the ideal lifestyle may be dependent on the
participant' s income and socioeconomic class. While direct evidence of a relationship between
the cultural domain of typical foods and income and socioeconomic class is not available, there
is strong evidence for a positive relationship between income, socioeconomic class and
household food expenditures (see Davis 1982 for a review) The lifestyle domain suffers from the
same lack of evidence, but the relation between the purchase of goods and income is well known
(Friedman 1957). While I was unable to control for all possible variables, the participants in all
the groups were selected for age and likely socioeconomic class. Income was not included as the
subj ect was difficult to speak about in both American and Mexican venues.
From my pre-dissertation fieldwork, I determined that the sample for the research should
be 18 to 40 years in low to lower middle class agricultural occupations. Table 4-2 lists the
demographic features of each dataset.
San Isidro Mazatepec, Mexico
I collected the data on Mexicans in the pueblo called San Isidro Mazatepec, or San Isidro
to locals. The town sits about 30 miles south-southeast of Guadalaj ara, Mexico in the district of
Jalisco. Approximately 1200 to 1500 people make the town their residence year round, most of
them farmers. The town revolved around the agriculture in the area. Tractors in the streets seems
to be ignored, one of the few specialized stores in the town sold seed and fertilizer. People talked
of weather patterns, rain, land and cattle. The farmers grow a plant known as blue agave for the
Tequila production in the region and maize and beans for more practical dishes.
The town grew out of a hacienda of the nineteenth century. The old hacienda mansion was
refurbished and along with its associated outbuildings forms the town hall and central plaza. The
only other features of the town are the soccer field, an arena for bull fights (toros) and a
cemetery. The cemetery easily topped the other attractions. It rested on a hill overlooking the
town as you entered. The graves were above ground tombs that were bleached white from the
sun and decorated with colorful flowers, religious symbols and miscellanea.
The streets are paved with small fist sized stones. The wear on them indicates that they
were laid down years ago and will not be replaced anytime soon. While they keep the road
passable when it rains, it is advisable to hold on tightly when driving through the streets.
Modern asphalt avoids the general area completely. Horses are as common to see on the street as
cars, although usually the vehicles were either pickup trucks or tractors. A number of times I
walked past a house to see a small corridor to the side. This led to an open area next to or behind
the house acting as a stable.
There are several small stores throughout the town, usually run out of the front room of a
house and are the only means of bringing common goods to the town. One store sells furniture
and another has enough business to specialize in farming supplies. Hotels or commercial housing
doesn't seem to be necessary. If you visit the town it is either for the day or you have relatives
with an available bedroom. A bus stop brings people from neighboring areas especially when
several large fiestas are thrown annually and draws crowds from Guadalaj ara.
While the town was supplied with running water and electricity, air conditioning was not
evident. This made little difference as the houses were constructed of large adobe bricks and then
covered in a plaster. Because the wall of the buildings were anywhere from fifteen to twenty-four
inches thick they served as a barrier to the elements. The sun would warm them during they day
then they would radiate the heat during the night keeping the house comfortable and be cooled
off for the following day.
The town did have a single school which consisted of one building, with five rooms. The
rooms divided the building into classes for different ages and on office on the end.
Northwest Kansas and Northeast Colorado, USA
I spent about two and a half months in northwest Kansas and northeast Colorado area to
collect data on Migrants and Americans. While there are three towns that I collected data in a
single description of the three is sufficient as they are very similar. Each town was approximately
thirty miles from the other in a straight line along interstate 70. Because the altitude is
approximately 3500 feet above sea level the area is know as the high plains. There is little foliage
in the area unless it has been cultivates. The weather patterns and lack of foliage create high
winds in the area. These are such a common occurrence that most residences will not notice a
wind less than 35 to 40 miles per hour. There is less than 15 inches of moisture per year which
classifies the area as a desert. However because of the rich soil in the area, along with irrigation
and conservation methods employed by the farmers the area is covered with fields of agricultural
grains. Wheat, corn, milo, soybeans and sunflowers are the primary crops. Wheat and corn are
the most significant.
The towns range from approximately 3500 people to 6500 people. All are dependent on
agriculture. Colby, the largest of the towns has a small community college that draws students
from the surrounding area. There is also some industry in the area. A small ethanol and biodeisel
plant is being built in Goodland. A few small feedlots survive in the area as well. All three towns
are dominated by businesses to support the agriculture. Each town has several implement shops
selling tractors, harvesters, farming equipment and their associated parts along with the standard
variety of discount, clothing and office supply stores. There are several banks in each town. One
peculiar element is found in Goodland, where Wal-Mart is the only grocery store in the county. It
came in the mid 1990's and drove out three of the locally owned groceries. This limits the
choices of products to only what they choose to stalk. It also means that most of the residents
will shop at the same store and have the same choice of products to buy. Colby has several
grocery stores and Burlington has a local grocery.
The migrant experience varies greatly between the towns. Burlington and Goodland are
more dependent on agriculture and the residents see migrants as necessary to the operation of the
farms in the area. Migrants have been used for labor for generations, with migrants making
repeated trips between Mexico and the US to work for the same family for many years.
Colby residents look less favorably on migrants. This is known throughout the migrant
community and little is done about the issue. The migrants will generally choose to live in
Goodland or Burlington which is friendlier. Colby residents have a stereotypical view of
migrants. I sat down with a number of teachers and administrators from the community college
and the topic came up as I described my research. They talked about the crime that migrants
cause, their lack of hygiene, overuse of social services (especially about issues in the school
system), refusal to learn English, and general lack of morals. These were less prevalent in
Goodland and Burlington. They were generally overridden by the recognition that without
migrants the local farmers would not be able to run their farms. Neither Americans nor Migrants
discussed the others lifestyle, pastimes or food choices. The exception came from Goodland
where one of the better restaurants in the area is run by a family that migrated from Chihuahua,
Mexico and serves a very authentic northern Mexican cuisine. The Americans appreciated the
quality of the food, even when they did not always know what they were eating.
Both Burlington and Goodland have several migrants who have settled in the area. Their
families have integrated into the community and started businesses or work in the commercial
establishments around the area. Some of these families provide support for the migrants as they
move in and out of the area. They often provide information about living arrangements, who
hires migrants and the best employers to work for. Each town had a few people who were heads
of these households and acted as a patriarch or matriarch for the Mexican community. All of
these individuals were well respected in both the migrant community and in the Anglo
community for their ability to place migrants in jobs and to keep disruptions in the migrant
community to a minimum.
While I was not able to see the living conditions in Burlington, I did have to opportunity to
visit the area where migrants stayed in Goodland. The area is a small street on the southern side
of the town. It consists solely of mobile homes. Each house can hold a number of migrants. It
was difficult to tell who lived where as people seemed to move between the houses frequently.
But they were little more than a place to sleep and take breakfast and supper. If a woman did not
have a job in the community they would often cook and clean for others in the house. Grocery
shopping was often done by both men and women. There was no grocery store in Goodland,
however the local Wal-mart had expanded to include food items. This was the only supply of
foods for the area unless a person drove to a surrounding town. Colby had a number of grocery
stores as well as general discount stores. Burlington was smaller and was limited to fast food and
a small grocery store.
Both men and women shop. Men seem to shop when women are not present to purchase
food and other supplies, but it is common to see them in the stores throughout the town. They
frequent the Mexican operated restaurants, but it is not uncommon to see them in the McDonalds
or Taco Johns. Perhaps the one place I did not see any migrants was in the local country club.
However an older Hispanic woman is the head chef for this establishment.
From interviews with Hispanic migrants and farmers in Sherman County there are
between 200-400 transnational migrants in Goodland and the surrounding area. Distrust of
outsiders in this population makes recruitment difficult, but the community is tightly knit,
making referral sampling a workable solution. Chain referral sampling, like snowball sampling,
uses referrals from participants to others in a community (Heckathorn 1997) but unlike snowball
sampling this strategy accesses multiple networks so that the sample is more representative of the
population under study. Chain referral sampling also resolves the ethical issue associated with
snowball sampling (where referrals by subj ects breach the privacy of the new participant without
the new participant' s consent), by using a locator, or participant who assists in the recruitment
procedure without disclosing personal information about the referred party to the researcher
(Penrod, et al. 2003). During the summer of 2005 I made contact with two migrants who are well
connected in the Hispanic community in Western Kansas. These two contacts agreed to act as
chain-referral locators for this research. Two additional contacts were made in Burlington,
Colorado and two in Colby, Kansas. These contacts not only allowed me access to the
community but assisted in the data collection as well.
When data are cultural and widely shared, sample size can be relatively small (Romney et
al. 1986). Handwerker and Wozniak (1997) argue that because cultural data are socially
constructed, sampling strategies for these types of data should focus on selecting informants who
have knowledge within the domain in question. This type of selected sampling is contrary to the
idea of random sampling and has implications for sample size in studies involving cultural
knowledge, namely that the number of respondents required to adequately understand a cultural
domain can be smaller than that needed for reliably estimating the prevalence of something in a
population, depending on the level of knowledge of the informants on a particular domain.
Because the domains chosen for this study are very common, all respondents should have
adequate knowledge to use a small sample (between 20-40 people).
The analysis for this research uses the cultural consensus domain (Romney, et al. 1986)
which not only demonstrates the degree of consensus within a sample but also estimates the so-
called cultural competence of each informant in a domain. The minimum number of participants
needed to describe a cultural domain is a function of three things: the cultural competence of
each participant, the confidence level required in a particular study, and the proportion of
questions asked that must be decisively classified (Romney, et al. 1986). This last point merits
some clarification. The larger the proportion of questions asked and the higher the acceptable
confidence level, the larger the number of participants. Cultural competence has the reverse
effect; the higher the competence, the lower the number of participants needed. Even in the most
extreme case of asking for a .999 confidence level and .99 proportion of questions correctly
classified, only 23 informants are needed if the average level of cultural competence is .60. At a
cultural competence level of .70 only 16 informants are needed (Romney, et al. 1986). Because
the competence of participants is unknown a sample between 30 and 40 should yield an
acceptable confidence level. Because of the difficulty of reaching the migrant population there
were only 27 migrant respondents, but, as I will show later, this sample was sufficient to test for
Table 4-2 lists the number of completed instruments by populations and type. Note that
the freelist and pilesort data were collected on each of the three domains. Note also that not all
of the collected data were used in the analysis. In some cases, participants did not understand the
data collection task (free listing or pile sorting). In these cases, a free list or a pile sort had to be
dropped from the analysis. However no participant' s entire set of data were dropped.
Dewalt and Dewalt (2002) argue that participant observation can enhance the quality of
the data collection and interpretation for most fieldwork situations. I used participant
observation to maximize my understanding of the cultural domains in this study. While I
participated in the activities of all three populations, each group was approached in a different
manner. In Mexico I spent the maj ority of my time participating in the cultural events and daily
activities of a Mexican family with whom I lived. My contact from San Isidro introduced me to
his family in the area. We spent a day talking with his relatives and seeing the town. After that I
traveled to San Isidro most days to meet with residents and observe their activities. During these
visits I did informal interviewing, speaking to people about various issues. Mexico/US migration
was a topic that I frequently used to strike up a conversation, and everyone had an opinion. I was
very open about my research and most people were glad to speak with me, although they found
my interest in foods and material goods a bit odd. I used these conversations to understand how
people thought of the structure of the domains in question.
Using the assistance of my Spanish instructor I spoke with the principal of the local high
school about hiring a few of the students to assist in the data collection. While they collected
some data for the proj ect their most important function was to give me access to their personal
networks. This was required to meet the requirements of chain referral sampling and it was also
necessary because of the short time period I had for field work.
To reach the appropriately comparable population among Anglo Americans, I spoke
with an administrator from the vocational school in Goodland and a professor at the community
college in Colby. Both allowed me access to students in their classes and agreed that the students
would meet the demographic criteria stated above. Data for this group were collected using two
assistants in each location who aided in recruitment and data collection.
The migrant population was approached with the help of a woman who had migrated as a
child and was currently working in the school system in Burlington, Colorado. She put me in
touch with an older woman who appeared to be a gatekeeper to the community. After speaking
with this gatekeeper several times she agreed to help and found two research assistants who
aided in recruitment and data collection of freelists and pilesorts. Again, their most valuable
function was to put me in contact with different networks of migrants. All of these individuals
were crucial to my research as the migrant population is very distrustful of any American at this
After the freelists were collected the two migrant assistants did not want to continue
working on the proj ect. One cited a lack of time and one didn't give a reason. I found out later
that she was not a legal migrant to the US. However, both assistants gave me the names of
friends who might be interested in working on thi s proj ect. The current political climate
certainly made recruitment of migrants as participant and as assistants very difficult.
Assistants and Training
At the end of the proj ect I had enlisted the help of 4 Mexicans in San Isidro, 2 Americans
in Colby and two Americans in Goodland, and 6 migrants living in either Goodland or
Burlington. All of these assistants were paid 100 to 200 dollars depending on their work load.
All of the assistants were trained to collect similar data. Each assistant was given a
practice questionnaire to complete. I explained the process just as I would to an informant. The
assistants were encouraged to ask questions and I tried to give some background to the
methodology. When there were finished they were asked to practice either with another assistant
or with me. They were also given some basic tips on follow-up questions and on how to address
any problems that might arise. For example, it is very common in pile sort tasks for informants to
ask what criterion the researcher wants them to use. The answer is always: "We want to know
what you think. There are no right or wrong answers."
After they had collected a couple of the questionnaires I contacted the student assistants
and asked to see the data. We discussed any problems or questions they had. After the freelists
were completed, they set up interviews for me to speak with some of the participants. Some of
the assistants were present for the interviews as we tended to do several in one sitting. They were
then trained to collect the pilesort data in the same manner.
All parts of the research were conducted in the native language of the participant. Mexican
and migrant participants created freelists of Spanish words. The interviews were conducted in
Spanish and the pilesort cards were written in Spanish as well. On a couple of occasions a
migrant participant would ask for clarification of a Spanish word, but this was infrequent and, I
believe, only occurred because of different usage patterns in a language and not from a real lack
of understanding. All parts of the research with American participants were conducted in
Freelists and Pilesorts
Researchers in cognitive anthropology have developed several systematic methods for
identifying cultural domains that have proven to be reliable. Freelisting develops a list of items
that a person considers part of a particular cognitive domain. A domain might be any element of
culture that has a common theme. Domains can be almost anything that people might group
together under a common term, for example: colors, plants, animals, illnesses, occupations, roles,
emotions, kinship, or even dirty words (Bernard 2002). During a freelisting exercise a
participant is asked to list the items that belong in that domain (Bernard 2002).
Following Dressler' s work in Brazil and Alabama in building domains of lifestyle and
social support (Dressler, et al. 1997; Dressler and Bindon 2000) I asked informants to free-list
items and activities specific to each domain. For the material goods domain, I asked that the
participant list all the goods a person owned when they have lived a good or successful life. The
other two domains were easier to explain and I only had to ask the participant to list typical foods
and typical pastimes.
Using the items collected from the freelist procedure a subset was selected for the pilesort
task to understand the structure of the domain. Item selection involves examining the frequency
and salience of items mentioned, and eliminating items that are not listed frequently. The exact
cut off point for accepting an item was determined by plotting the frequency of the items listed.
Typically, these scree plots show a steep initial decline, followed by a gentler slope for several
items, and culminating in a long tail of items that appear just once (Borgatti 1996). A typical cut
off point is the elbow where the slope becomes less steep. In the example in Figure 4-2 the most
likely cut off point would be when an item was mentioned more than four times. Table 4-3 gives
this cut off point for each of the freelists.
Pile sorting a subset of the responses from the freelists provides data on the cognitive
structure of a cultural domain. Stacking the cards in the same pile indicates that the participant
sees some sort of similarity between the items. With n items in a pile sort, the informant makes
n(n-1)/2 decisions as he or she places the cards in piles of things that go together (Burton and
Two types of pilesorts were used here. In the open pilesort, the participant is handed a
stack of index cards with the name of one domain item written on one side and an identifying
number written on the other. Participants are asked to sort the cards into different piles, or stacks,
by placing similar items together and dissimilar items in separate piles. They are told that they
may have piles with only one card, but they cannot make all of their piles from single cards and
they cannot have only one pile of all the cards. As each participant completed the task, either I or
my assistant recorded the numbers on the back of the cards and continued with the other
The second pilesort--known as a constrained pilesort--tests which items informants
believe belong to a domain. In this procedure participants are asked to separate the cards into
those that belong to the domain and to discard the rest. This procedure is a proxy for a true-false
test in which people are asked 'should this item be a part of this domain?' (Weller personal
communication). The corresponding numbers were recorded and the participant continued to the
Between completing the freelists and pilesorts a small sample of participants were
selected to participate in personal interviews. These interviews were informal and consisted of
questioning the participant about the domain items and why they may have been included, as
well as how they would classify these items. Because of the informality of the process there was
no hard format to the questions. What I was trying to understand were the characteristics of the
items found in the freelists that put them in certain groups. These interviews were only partially
recorded as a number of people asked that the recorder not be used. Notes were taken during all
of the interviews.
Basic demographic information was collected on all participants who completed the
freelist or pilesort sections, including: age, gender, marital statue, number of children, number of
years of education, profession and city born. Table 4-4 summarizes the all of the demographic
variables except profession and city born.
Cohen's Transnational Score
Transnationalism was measured to understand the relationship between transnationalism
and the differences between the cultural knowledge of migrants and non-migrants. A problem
with measuring the degree of transnationalism of migrants is the different definitions of the
phenomenon (Cohen, et al. 2003; Guarnizo, et al. 2003; Irzigsohn and Saucedo 2002). There are
varieties of transnationalism that include social, economic and political behaviors and that
involve a number of different activities (Portes 1999). Itzigsohn and Saucedo (2002) asked
respondents to rate their participation in hometown associations, their sending money home for
community proj ects, traveling to public festivities, participating in local sports clubs, and
participating in charity organizations that are linked to their country of origin. A survey of nearly
2000 migrants in Vancouver, Canada, asked about migrants' friends and family, how frequently
(and with what method) they stay in touch with friends and family, the frequency of travel to the
home country, ownership of either a business or property at home, and whether they provide
financial assistance to family and friends who still reside in the home country (Hiebert and Ley
2003). Cohen et al.(2003) conducted a survey in the State of Oaxaca, Mexico that measures ties
to previous migrants (MIGFRIEN), how migrants paid for their trip (USCOSTS), where they
lived once they arrived in the U. S.(USSTAY), the number of years the migrant has remitted
funds (REMTOT) and the number of migrants the household had previously sent (TOTUS).
These measurements focus on the migrants use of kinship ties to aid in the migration process and
the use of remittances to maintain a strong relationship with the sending community (Cohen, et
al. 2003). Because my research is concerned with the degree of transnationalism and how it
might relate to other variables, l used Cohen's work as a blueprint for my questionnaire. The full
questionnaire can be seen in Appendix A. The transnational data are shown in table 4-5. The
transnational score is calculated with the following formula: Cohen's individual transnational
score (CITS) = MIGFRIEN + US COSTS + US STAY + (REMTOT/TOTUS).
The individual transnational score is calculated with the following formula: TRANSTOT
= MIGFRIEN + USCOSTS + US STAY + (REMTOT/TOTUS). The average TRANSTOT score
is 2.83, standard deviation of 1.75, minimum of 1 and maximum of 7. The mean score is very
close to Cohen's scores of Oaxaca transnational migrants (2.9). Cohen's data showed a larger
maximum (25) and more variance (sd 3.5). However, Cohen's data comprise interviews with 323
people in 11 counties in Mexico.
Acculturation was measured to understand how it is related to the changes in cultural
knowledge of migrants. The measurement of acculturation is less problematic than measuring
transnationalism as several instruments have been developed and tested in many research
proj ects (see Zane and Mak 2002 for a review). The Acculturation Rating Scale for Mexican
Americans Revised (ARSMA-II) (Cuellar, et al. 1995) is a Likert type scale that measures
language, ethnic identity and ethnic interaction as factors in the process of acculturation. The
method is multidimensional and measures the participants' Anglo orientation (AOS) as well as
their Mexican orientation (MOS). It can thus be scored as a series of linear outcomes or
multidimensionally. Linear scoring places the migrant in one of 5 categories of acculturation
(from unacculturated to fully assimilated) by subtracting their MOS from their AOS. These are
single scores that can be compared to others. Orthogonal scoring places the migrant into one of
four categories based on high or low AOS and high or low MOS: Traditionals (low AOS, high
MOS), Low Biculturals (Marginalized), High Biculturals, and Assimilated (low MOS, high
AOS), based on where they are in their extent of involvement in Mexican/Anglo culture and the
acceptance of the attitudes and behaviors in Mexican /Anglo culture. Table 4-6 shows the
numerical and categorical scores for the Mexican migrants in the study.
of the domain
the structure of
Figure 4-1 Overview of research design
Table 4-1 Number of responses collected by type and population
Type of Data Mexicans Americans
Freelists (on each domain) 27
Demographic Information 27
Cohen's Transnational Test
Open Pilesorts (on each domain)
Closed Pilesorts (on each domain)
Cohen's Transnational Test
Table 4-2 Number of items selected from freelists
Mexicans Americans Migrants
Foods 23 22 27
Passtimes 28 36 24
Material Goods 26 30 24
Table 4-3 Participant demographic information
Average % Number Ave Years
Age % Female Married Children Education
Mexican 32.05 0.70 0.45 1.03 10.13
Migrant 31.37 0.56 0.78 1.93 9.22
American 23.43 0.47 0.20 0.47 13.59
Mexican 36.21 0.66 0.39 2.10 10.52
Migrant 32.07 0.68 0.86 1.53 9.75
American 22.33 0.57 0.10 0.37 12.93
Figure 4-2 Frequency of freelist items for migrants
Migrant Things from a good life
Table 4-4 Cohen individual transnational scores
USCOSTS REMTOT USSTAY TOTUS
Table 4-5 ARSMA-II scorl
There are two goals in this research. The first is to understand how the elements of the
cultural domains differ between migrants and non migrants. The second is to examine the
structure of the domains and how this may change because of migration behavior. This chapter
answers the first question while the next chapter answers the second one. To focus this analysis I
H1: The cultural domains of transnational migrants are a mix of elements from the home
and host society's cultural domains.
Alternative H1: The cultural domains of transnational migrants will have elements that
are neither in the home nor host country's cultural domains.
l use data from the free lists and the constrained pilesort task to test for consensus in the
content of cultural domains. I use two datasets in an effort to overcome the critique that cognitive
anthropology in general, and cultural consensus analysis in particular do not focus sufficiently on
variation in cultural domains (Aunger 1999; Aunger 2004). Aunger (1999) argues that it is not
the frequency with which items are found in a culture that makes them cultural. Rather, it is the
fact that they are transmitted to others. Aunger' s focus on transmission allows a much broader
range of items to be included as cultural instead of only those that are common. Using both
datasets allows me to show the similarity between the items as well as the variability within the
domain. Using the freelisted data also allows me to examine the elements of the domain to
understand its composition.
The chapter begins by describing the cleaning and renaming of the datasets. A qualitative
analysis of the common elements of the dataset is then examined. The composition of the
domains is further explored with the freelisted data by analyzing the overlap of items in each
domain. This analysis is done with both the full freelists and a reduced freelist for reasons to be
Recall that three types of data (freelists, open pilesorts, constrained pilesorts) were
collected from three different samples -Americans from northwest Kansas, Mexicans from San
Isidro Mazatepec, Mexico and transnational migrants who have moved from Mexico to the
northwest Kansas. Tables 5 -1, 5-2 and 5-3 list the number of usable responses for each sample,
domain and instrument type. The numbers in the cells may vary because a respondent refused to
answer a particular section or because they produced unusable data. The latter error occurred
when my assistants were being trained in the data collection method.
Domain and Population Names
Because this analysis will deal with three populations and three cultural domains a
shortened title will be given to each group and domain. "Nonmigrant Mexicans" will be called
Mexicans; "Anglo Americans" will be called Americans; and "Transnational Mexican migrants"
will be called Migrants. The domain of "things that people own when they have had a successful
life" will be called Things; "typical pastimes" will be called Pastimes; and "typical foods" will
be called Foods.
Cleaning the Freelists
After the raw freelists were collected and entered verbatim into an excel sheet, each list
was cleaned. Cleaning the datasets consisted of making a number of modifications to particular
entries to produce consistent data for the analysis, taking care that the modifications retain the
intent of the respondent. Fixing errors in spelling is a low-inference modification. Synonyms
were combined so that that 'cattle' and 'cows' were placed together' and phrases like 'going for
a walk' and 'walking', or 'eating out' and 'restaurants' were combined. Modifiers were removed
so that 'family' and 'happy family' were combined into one category in the domain of Things.
These are slightly stronger-inference modifications.
There are some situations where similar responses were not grouped. This occurred when
people mentioned different levels of a domain. This was a particular problem in the domain of
foods where people would mention 'chicharron' as well as 'chicharron con chile' or 'chicken'
and 'chicken dumplings'. These items were not combined into a single category because they
were, at times, mentioned by the same respondent and because I believe that they intended two
separate and different items. This choice is the result of carefully examining the data and
comparing the items to my experience with the informants. I believe that combining items in
these cases would result in a critical loss of variability in the domain.
Large parts of any freelist are items listed by a small number of people or by a single
person. Table 5-4 lists the items that were listed by at least 20% of the respondents.
The most noticeable feature of Things is that 'car' and 'house' were the top two items in
all three domains, indicating a universal sign of success across the cultures. 'Clothing' and
'businesses' were also mentioned by all three groups and may be part of this universal pattern for
these cultures. I expected these items to be listed by all of the groups as well as the more
common household items. However there was considerable variance on the particular items that
were listed besides 'car', 'house', 'clothing', and 'businesses'.
Migrant women listed clothing and j ewelry almost exclusively. There was one man that
listed clothes and another that listed jewelry, indicating that there is a gender bias on the items
that are listed. 'Money' was listed by Mexicans and Migrants, but not by Americans. This
interesting divergence is probably a response effect. In English the question was phrased as,
"Please list the goods or possessions that people own when they live a good or successful life."
In Spanish the question was, "Por favor, anote las cosas que una persona puede conseguir cuando
tiene una buena vida, o sea, cuando la persona tiene much exito." The English version
emphasized the idea that the question was about material goods by using the words 'goods' and
'possessions'. On the other hand, such insubstantial things like 'friends' and 'family' were only
listed by one or two of the respondents in each group. They did understand that it was material
goods to some extent, but the domains are interpreted different for each group.
'Travel' was only listed by Migrants. This may indicate that migrants may have a
propensity for traveling or they may have learned to enj oy their travels abroad. Migrants also
listed 'computers,' 'cell phones,' and 'fumiture.' These items were shared by Americans but not
by Mexicans indicating that migrants may have picked up a taste for electronics and home
furnishings in their travels. Americans used lower levels of contrast in listing Things, naming
brands when they mentioned 'dishwasher,' 'dryer,' and 'washer.' This may imply that the
domain of Things is more defined in American culture.
Recreational vehicles were listed only by Americans. These items may be large enough
purchases to be out of the realm of possibility for typical Mexicans and Migrants or in the case of
pastimes Migrants may not be aware of the usable recreational facilities. Only Americans listed
'land' as a material indicator of a good life. Neither Migrants nor Mexicans mentioned this item.
This finding seemed odd, since many Mexican migrants are known to purchase land in their
home communities with funds acquired abroad. I looked for subgroups within the American
freelists that listed 'land' and while there is not much variability in the dataset overall, the
respondents who listed land showed no consistent pattern in age, gender, marital status, number
of children, years of education or profession. Another possibility may be that because Mexicans
use the money they make in the U.S. to build housing in Mexico, land may be included in the
idea of house, which was listed by all groups.
Table 5 -5 lists the Foods there were mentioned by more than 20% of the informants.
Similar to Things, Foods has some universal elements like tacos and burritos (clearly from the
Mexican domain) as well as items like 'hamburguesa' espagueti and ensalada. These last items
are more international but have long been mainstays in the Mexican diet. In one instance, the
respondent used the English word, 'spaghetti' instead of the Spanish "espagueti'. Chile colorado
is a commonly listed food among Migrants, but not among Mexicans. It appears to be a Mexican
dish that is adapted to local tastes and to local ingredients. A lot of what is today standard fare in
Tex-Mex cuisine--things like chile con carne and fajitas--began this way, as did many standard
dishes (like chop suey) in Chinese American cuisine.
Table 5 -6 lists the Pastimes mentioned by more than 20% of informants. Similar to the
other domain there is a small set of Pastimes present in all three of the samples, including the
activities of watching TV and reading. 'Listening to music' is prominent in the lists of Migrants
and Mexicans but is absent in the Americans' lists, as were a number of other items. With the
popularity of Ipods and other portable music devices, music may not be considered an activity
that is done as much as it is simply an accessory like wallet or purse that accompanies the person
almost continuously. Two items that stand out in this respect are the activities with friends and
family, present only in the lists from Mexicans and Migrants. In fact there were very few items
that Migrants have incorporated into their domain of Pastimes. One exception is 'deportes' or
'sports' that is present in Migrants' lists. While the generic term 'sports' does not appear in the
Americans' freelists, their lists are dominated by the names of specific, organized sports.
Another shared activity by Americans and Migrants is 'shopping'. Americans included
items like surfing the web and playing computer games. Neither of these was present in the other
groups' lists. Even though the sample of Americans was mostly young, single adults there was
little mention of social activities like going to bars, dancing or drinking. In general the activities
of Migrants not only include more social activities but also more informal activities. These
activities, like 'taking a walk' or 'going to the park', do not require any equipment or facilities.
They also do not include 'toros' (going to the bullfights) and 'futbol' (soccer), probably for the
same reason lack of facilities.
Domain Overlap Analysis I
The overlap in groups can be examined with Venn diagrams or with correspondence
analysis. Correspondence analysis uses the frequency of items mentioned by all groups to create
a plot with both the items and samples plotted on the same axes. The distance between the
groups can be measured directly. The Venn diagram does not use the frequency of the items
mentioned but is simply based on whether an item is part of the group or not. While the
correspondence analysis uses the additional frequency information it can become incoherent
when a large number of items are present in the data, as they are with the domains under analysis
here. Venn diagrams are used for clarity and ease of interpretation. The software used to create
the diagrams shown here is freeware called Venn Diagram Plotter (Littlefield 2004).
Venn diagrams are an area proportionate graph where the area represents the number of
items in each group. The area of the region where the boxes overlap represents the number of
items that occurred in at least two domains. For example, in Pastimes, dancing was mentioned
by Migrants and Mexicans but not by Americans. Therefore, it would be counted as one item in
this overlap region. The center areas, where all three boxes overlap represent the number of