This paper will be presented at the PSA Annual Conference, Glasgow, April 9-12th, 1996 at Political Science Specialist Group panel on "New Boundaries; New Identities". It will also be printed in the Conference Proceedings.

Introduction

The ugliness of generations as a tool for analysing politics is easily perceived; for defining a generation is thoroughly vexing, and any such definition defies easy operationalisation. I argue in this paper that one reason why generations are so difficult to define and operationalise has to do with time. Accordingly, I argue that political scientists have tended to be more interested in 'timeless' — and therefore less confounding — aggregates of people such as: nations, genders, classes, interest groups, sects, etceteras. These groups are 'timeless' in the sense that understanding both externally (quantitatively) measured and internally (qualitatively) experienced time is not essential to understanding them. And, as I explain, this is fundamentally different from generations for which understanding both kinds of time is crucial.

The ugliness of generations is made most evident when they are put to use by political scientists. Given the complexity involved in understanding both internal and external time vis-à-vis generations, it is not surprising that errors in application usually occur at the most basic level. To take one notable example, Inglehart in his analysis of intergenerational, postmaterialist value changes confuses his understanding of generations by muddying the important differences between generations and cohorts. 2 This sort of confusion of terms indicates a fundamental misunderstanding of generations. I contend in this paper that for generations to be useful to political scientists, they must be separated from seemingly similar divisions of time such as cohort and decade, as well as from other social aggregates such as class and nation.

But this is not to suggest that generations are necessarily as immediately intelligible as other divisions of time and other social aggregates. Indeed, while this paper begins by introducing some transparent observations about generations, the discussion rapidly unfolds into something more complicated. Hence the "problem of generations" 3: while we feel confident in using generations to carve human biological and socio-historical reality into recognisable units, upon closer scrutiny our confidence wanes. While we tend to think and talk quite comfortably about generations, that comfort often obscures the reality that we are not entirely sure what, in the first place, a generation is and, more fundamentally, whether generations are a relevant analytic tool.

In this paper, I approach the use of generations in political science in five steps. First: I analyse the "problem of generations" as a means of introducing some of the basic elements of a generation. Second: I compare generations and other social aggregates leading to a sketched definition. This part of the paper also introduces the idea that both internally experienced and externally measured time is the essential element to defining a generation. Third: I discuss, accordingly, the importance of taking time seriously with respect to generations. Fourth: I detail some examples of generational analyses in political science. These examples, I argue, highlight the most common, underlying mistake committed by political scientists when analysing generations: namely, inadequately accounting for both quantitative and qualitative time. Fifth: In conclusion, I indicate the form a further analysis of "generations" in political science should take.

Before introducing the "problem of generations", I would like to indicate that the broader purpose of this paper is to lay the groundwork for conceptualising and operationalising generations. With an operationalisation in hand, I hope ultimately to test my understanding of generations by analysing certain political behaviour of certain generations. 4

The "problem of generations"

The central and indisputable meaning of generations has to do with the biological, filial progression from parents to children: generations differ on the basis of filial relationships. This most basic conclusion implies that a generation cannot be shorter than it takes for people to become sexually mature: that is, the age at which children can become parents. This simple insight is as obviously sensible as it is easily ignored. 6

We also understand generations as differentiating themselves through a process of conflict with each other, pitting old against young. For this reason, children's personalities can develop in reaction to the personalities of their parents: "the personality of the son is often directly opposed to that of the father; spoiled children are likely to become tyrannical parents; a conservative and intolerant king is often followed by a progressive and liberal crown prince, and so forth". 7

That generational differences arise merely as a result of a conflict between generations of parents and their children is an overly simplistic account; for at the same time are always more than two parent-child generations. In modern society, there is no one generation of parents and one generation of its children. In other words, generations in modern society, while fundamentally based on the differences between parents and children, must be understood as a social and not strictly familial phenomenon.

A more sophisticated reason for generational differentiation has to do with the interplay of old and young generations. This interplay need not be restricted to generations of parents and generations of children. Between, before and after any two parent-child generations can be members other generations whose influence must be taken into account. For example, according to one commentator, in late eighteenth-century America the disagreement between some prominent Antifederalists and Federalists was between members representing two different generations, although not literally filial generations. 8 Patrick Henry, for instance, one of the Antifederalist leaders, was only six years younger than John Jay, one of his Federalist counterparts. An ancient Arab proverb is helpful in at least partially explaining why this and other non-filial generations emerge: "Men resemble the times more than they do their fathers". 9

Indeed, many not necessarily filial generations exist at the same time. The resulting interaction of generations creates a complex stacking of influence and counter influence. And this multi-levelled reciprocal relationship indicates one reason why generations do not suddenly start and end: namely, the cross-fertilisation of generational attributes and ideas. The relationship between poet Shelley and philosopher Godwin illustrates the point. "It would not be an exaggeration", explains Brailsford, "to say that Godwin formed Shelley's mind". 10 Godwin's ideas obviously held currency for Shelley, despite the thirty-six year age difference between them. For Godwin and Shelley, at least, their respective generations were not hermetically sealed from each other.

This variety of insights into the character of generations indicates why much confusion surrounds the scholarly and popular use of "generation". As a prerequisite to examining the source of this confusion, I will compare generations and more recognised social aggregates. By so doing, I will not only lead up to a discussion of time and generations but sketch a definition of "generation" as well.

Generations are a unique variety of social aggregate

A community group's unity evolves naturally and spontaneously, creating and building upon a spiritual connection between its members. For that connection to develop in the first place, members of a community group "must have concrete knowledge of each other", and to maintain that connection, they must also be in close physical proximity. 12 Association groups, on the other hand, are consciously created to achieve a specific purpose. Therefore, an association group will usually have written statutes stipulating why the group exists in the first place — its purpose — and how it can be dissolved should its members see a schism between the group's purpose and actions. This document unites members of an association group in the absence of the spiritual and proximal connections of a community group.

In contrast to community and association groups are two similar kinds of social aggregates, socio-economic class and generations. I will discuss each in turn. Mannheim argues that "class position can be defined as the common 'location' ... certain individuals hold in the economic and power structure of a given society as their 'lot'". 13 Most have an insight into their own socio-economic class both absolutely and relatively. And although we can move between socio-economic classes, as a result of any one of a variety of factors, that class continues to exist for those who remain.

Socio-economic class is more obviously distinguishable from association groups than from community groups. Unlike an association group, a socio-economic class is not formed by a conscious act of founding and cannot be dissolved by any procedural method. Like members of a community group, members of a socio-economic class are held together by a spiritual interconnection: a common feeling that they share a "'lot'" in the "economic and power structure of a given society". This shared feeling explains how the 'community' of members of a socio-economic class are held together even though they will not all be in close physical proximity, something that would make a community group impossible. Thus, a socio-economic class can develop a spiritual intraconnection that binds its members together at least to the extent that they feel part of one class as opposed to another.

Similarly, members of a generation feel a sense of commonalty as a consequence of their generational 'lot'. Generations, however, differ from socio-economic class in two important ways, the explanation of which will flesh out what I mean by 'lot' with respect to generations. Firstly, an individual cannot move within or between generations. As Wolfe puts it, "you belong to [a generation] .... You came along at the same time. You can't get away from it. You're a part of it whether you want to be or not". 14

The second and more significant difference between socio-economic class and generations has to do with time. While both types of aggregates demarcate certain coordinates in society, a generation's coordinates exist on two planes: the planes of biological and socio-historical time. A generation is a group of like-aged individuals who are commonly imprinted by socio-historical events because they experience those events at a similar age (and youth is the stage of life during which we are most imprintable 15). Once imprinted in this manner, members of a generation are more likely to interpret future socio-historic events similarly.

Fully explaining this sketched definition would go beyond the purpose of this paper. 16 Suffice it to say here that at the most basic level, generations have to do with socio-historic events like-aged people encounter together. This common encountering, because members of a generation experience it at the same or similar age, imprint upon them similarly. In turn, this common imprinting leads them to interpret future events similarly in a self-reinforcing cycle of experiencing and interpreting. This, in short, explains why we commonly (and largely correctly) regard a generation as being comprised of like-aged individuals who interpret at least some aspects of reality similarly, which is to say they manifest a generational 'attitude' or 'personality'.

Thus, to return to the comparison of socio-economic class and generations, I conclude that generations blend experiential time, the first set of coordinates, and biological time, the second set of coordinates. Biological time is unimportant to socio-economic class, let alone the combination of biological and experiential time.

Generations must also be considered with respect to physical space in so far as a similar history can usually only be experienced by individuals living in a similar place. Of course duke and dustman who belong to the same generation may interpret socio-historical events differently, even though one may live upstairs and the other downstairs in the same house. Generational commonalties are never manifested by all members of a particular generation. Rather, generational attitudes and behaviours are exhibited by most members of a generation most of the time.

I should also point out that the distinction between generations, on the one hand, and association and community groups, on the other hand, fails to clarify certain crucial phenomena in politics. For example, a nation has two significant similarities with generations: both are cross-class aggregates of people who share a 'commonalty of fate' or 'shared destiny'. Despite these significant similarities, the essential difference between generations and nations, as well as other groups studied by political science, is the way in which time is accounted. While nations and generations have much in common, nations are not built upon the combination of both biological and experiential time in their formation and understanding.

The importance of taking time seriously

Since Plato, analysts of human behaviour have been largely concerned with discovering universal principles and properties. In this pursuit, time can be understood only causally: if ideas are timeless, time can be nothing more than a progression from a theoretical 'Time 1' to 'Time 2'. Time, in this context, does not matter except to the extent that it explains the objective 'distance' between a cause and its effect(s). In nature, time is causal and moves in cycles: an acorn drops to the ground; it develops into an oak tree; the tree lives, dies and rots; the rotted tree provides nutrients for future oak trees. Time so understood can be described as natural, causal, biological, cyclical, scientific, quantitative, external, or objective. Of course cyclical time is not circular time; for the same oak tree never reappears. The progression from birth to death to birth is abstractly cyclical in so far as it keeps time with an externally measurable and reasonably predictable natural rhythm.

Hegel argues that a cyclical conception of time cannot adequately explain human reality. He points out that while nature moves in cycles, human interpretation of reality does not. 17 Thus, while an acorn is like a child in that both will, with the right conditions, grow, reproduce and die, humans also experience time. Time so understood can be described as linear, subjective, internal, qualitative, historical, or experiential. A human life is therefore both cyclical — we are formed from ash and will return to ash — and linear — our lives are narratives that we create as we interpret events. The most important narrative of Western civilisation is the linear history of Christianity. There is a beginning: Adam and Eve in Eden. There is a middle: what has and has yet to come after the Fall. There is an end: the second coming of Jesus Christ. Implicit in any narrative is a subjective, qualitative accounting of time. We push and pull and bend and stretch time when making stories. That 'a watched kettle never boils' demonstrates that we are able to interpret time subjectively, despite the objective reality of seconds, minutes and hours.

Historical narratives defy cyclical classification because this sort of classification can only admit of cyclical time whereas humans interpret time linearly, with respect to specific circumstances. Thus, "to respond to a particular situation, event, or state of affairs is not to respond to any situation, event, or state of affairs with the same or similar properties in some respect, it is to respond to that situation conceived by both the agents who respond to it and those whose actions constitute it as particular". 18 In other words, human existence cannot be separated from the instantaneous characteristics that particularise that existence. Reality cannot be adequately reduced into a set of intersecting universalised properties, something that quantitative analyses attempt to achieve.

The use of generations in political science

Typically, analyses of generations and politics takes form in studies of the relationships between certain measurable attitudes and behaviours, on the one hand, and equally measurable cohort effects, as opposed to ageing (life-cycle) or period effects, or both, on the other hand. 19 Like the prominent political scientist Inglehart who measures intergenerational value changes, many who discuss generations are unsure about what is and is not a generation. This lack of certainty is evinced most clearly and commonly in the interchanging of "cohort" and "generation". Inglehart slips from "cohort" to "generation" without specifying the differences between them. 20

Similarly, when discussing generations and political apathy in America, Bennett says that: "Less certain is whether the most recent cohorts have been especially apathetic or whether those Americans born from 1940 to 1947 … were an unusual generation". 21 How can a generation be only seven years long? What is the difference between a generation and a cohort? Why would a generation/cohort begin in 1940 and not, say, in 1941? Answering these basic — and difficult — questions is possible, but Bennett does not make the attempt. Nor could Inglehart's writings help in answering these questions. Bennett and Inglehart are but two examples of many. Thus, it is often the case that generations are used as units of comparison by analysts who are not themselves clear about what constitutes a generation. Discussing their "generations", then, can be confusing.

As I have indicated, perhaps the most common confusion when discussing generations is the mixed usage of cohort and generation. Is a generation the same thing as a cohort? Glenn, in addition to Inglehart and Bennett, talks as if they were. A cohort, he says, is a group of people within a specifiable population that has experienced the same significant event(s) within a certain period of time. 22 His definition is unusual and includes some of what I put forward earlier as essential components of a generation. Most others, including Inglehart and Bennett, however, use cohorts merely to describe a group of individuals delimited by mathematically compartmentalised blocks of time, as in "Age less than 35", "Ages 35-49" and "Age 50 and over". 23

In order to avoid fundamental confusion, I recommend that analysts of generations avoid the term cohort whenever possible for two reasons: because cohorts usually refer to discrete, easily-definable sub-populations and generations do not; and because cohorts usually pertain to biological time only, whereas generations also have to do with experiential time. 24

Preference for cohort analyses implies a concern that analyses of generations over-generalise. However, dividing generations into cohorts invites further divisions to achieve yet more precision which, after all, is the point of preferring cohorts to generations. Of course my second objection — that they invite finer and finer distinctions — could also be levelled against those who discuss intragenerational divisions, such as "Generation Y", the members of which are said to be born between 1979 and 1987, within Generation X. 25 To avoid being sucked into a bog of generational subdivisions, and subdivisions of those subdivisions, a bog made all the more gooey by popularised generational analyses, the analyst must be strict and clear about what is meant by a generation. At the centre of mainstream social commentary, threatening to corrupt our understanding of generations, journalists have created a plethora of 'generations' that can potentially be useful only in describing parts of generations. "Baby Buster Generation", "Slacker Generation", "Can-Do Generation", "Generation Y", "Me Generation", and "Protest Generation" is but a small selection. 26

Thus, beyond the confusion arising because of the mixed usage of cohort and generation are the many subdivisions of generations and cohorts. Even more confusing, the concept of generations is often not clearly distinguished from other divisions of time, such as the very popular decade, leading to the creation of entities as dizzyingly incoherent as Jennings and Niemi's "youth generation [at the centre of] … the protest cohorts of the late 1960s".

Conclusion

Ultimately, a reasonable approach to the study of generations must combine both vital factors (biological, cyclical time) and socio-historical events (historical, linear time). This being said, however, there may not be a generational rhythm in society, and even if there were, it may not be discernible.

To understand generations is to understand the rhythm of life, from birth to death. In an obvious sense, everything we do and think begins with our biology, but this should not lead to an over-reliance on vital factors, for a thick layer of human thoughts, actions and interactions exists above biology. While this is a reasonable conclusion, many difficult questions must remain unanswered in this paper: How does one determine the beginning and end of a generation? How can subjectively experienced time be operationalised? How can the varying importance of socio-historic events be measured let alone identified?

This paper cannot answer these questions. In other words, it cannot put forth a methodology by which one can demonstrate that generations ought to be included among race, gender, class, and interest groups as a useful measure of political analysis. This paper has merely pointed out that analysing generations is complicated primarily in so far as generations live somewhere between cyclical and linear time. As a consequence, when political scientists talk about generations they usually confuse their efforts, talking not about generations directly but only potentially indirectly.

To promote generations as a useful tool of political analysis, I will need to conceptualise "generation". I will then need to operationalise that conceptualisation, and ultimately test that operationalisation with data. To demonstrate that an adequate conceptualisation of "generation" can be operationalised and successfully tested will go a long way towards arguing for taking time seriously in political science through generational analysis.

Notes and References