Defining Leadership

Abstract

This essay examines the concept of leadership as it is commonly understood within the field of leadership studies today. The inquiry is framed by an analysis of three generally accepted definitions of leadership. I look at the selected definitions from four angles, which I call the four dimensions of leadership: the behavioral (what the leader does, or ought to do, that makes it leadership), the asymmetrical (in what sense a leader is different from the others in the group), the social (what it is that the leader is leading), and the teleological (what the direction is). By doing this, I find that these definitions can be problematic, if instructively so. I identify two underlying concerns. The first problem is the tendency to fuse informal and formal leadership into a single concept. I argue that informal and formal leadership can be more helpfully understood, rather, as two separate, if sometimes overlapping, concepts, each needing its own definition and analysis. This is because formal leadership is a functional concept, while informal leadership is not; this entails not only that each type of leader is leading a different type of entity, but also that each is judged by different standards. The second problem is, paradoxically, a bias against leadership. To put this less polemically: scholars who recognize that organizational success can be hindered by excessive emphasis upon hierarchy, and who thus study the dynamics of egalitarian collaboration, can inadvertently generate confusion by allowing the honorific of “leadership” to be applied also to that work. So, I propose definitions of the two key concepts, formal and informal leadership, with the suggestion that this provides a helpful method of sorting leadership scholarship into three categories: formal leadership; informal leadership; and egalitarianism.

The Conceptual Clutter

For conversations to be productive, we need to be talking about the same thing. If I sit expectantly at Louie’s Dive Bar while you wait for me at Louie on DeMun, we may regret not having taken more care with our definitions. Leadership scholars, correspondingly, often note the lack of an agreed-upon definition of leadership, and, correspondingly, lament that researchers have ended up in different places.

In the middle of the last century, Launor Carter complained: “Leadership behaviors are any behaviors the experimenter wishes to designate or, more generally, any behaviors which experts in this area wish to consider as leader behaviors” (Carter 1958, 24). The product of this permissiveness, noted by Thomas Cohn and D. G. Browne in their introduction to the volume containing Carter’s essay, is: "…the leadership literature is a mass of content without any coagulating substance to bring it together or to produce coordination and point out interrelationships" (Browne and Cohen 1958, iii). Have things improved? Martin Blom and Mats Alvesson recently assessed the current scene with, in effect, a paraphrase of that decades-old indictment: “The academic leadership literature…refers to leadership as everything and nothing.” The result is that it is little more than “noise and confusion created by the big, loud, and sometimes disharmonious leadership chorus” (Blom and Alvesson 2015, 480).

These are not voices crying in the wilderness. Even more recently, John Antonakis and David Day express misgivings about whether we can “get leadership scholars ever to agree on a definition.” And they colorfully assess the resulting situation by asking us to “imagine taking pieces of several sets of jigsaw puzzles, mixing them, and then asking someone to put the pieces together into one cohesive picture” (Antonakis and Day 2018, 29). They are hopeful that things are improving; but if the pieces continue to be fabricated from multiple mystery puzzles, we may wish to reserve judgment.

Why Start with Textbook Definitions?

The first challenge is to produce a reasonable account of what leadership scholars have in mind when they examine leadership. Different scholars, no doubt, have different things in mind, but it can be a fair and helpful strategy to seek out the most generally accepted definitions in the scholarly literature. There is no clear candidate, however, for a prevailing or even an influential definition. There have been a handful of interesting attempts in recent decades, but none has either established a foothold or generated much interest in the definitional project.¹ Richard Barker writes, “Not defining leadership seems to be an accepted practice among scholars who discuss leadership” (Barker 1997, 334). Barbara Kellerman confirms this, at least in her own case: “I avoid like the plague definitions of leadership (of which, at last count, there were some fifteen hundred) and theories of leadership (of which there are around forty)” (Kellerman 2012, 11). Keith Grint goes so far as to speculate that leadership cannot be defined, proposing that it is an “essentially contested concept” (Grint 2005, 1).²

Noting that the English phrase “a textbook definition” designates a particularly current and clear statement of an idea, my solution is to focus upon the definitions that are offered in three of the most up-to-date and widely used leadership studies textbooks. I cannot claim that anyone beyond the authors of these volumes adheres to their definitions. In the absence of reasonable alternatives, however, it seems productive to take their views as a proxy for prevailing assumptions of the discipline. These comprehensive volumes survey the state of the academic field of leadership studies for the purpose of preparing future leaders and future leadership scholars.³ In particular, each of them begins with a review of the decades-long effort to define leadership, and concludes with its own definition that is designed to fit current academic usage (each of them, as it turns out, differing from the other two only in superficial ways). They are authored by experts who have contributed other respected books and articles to the field, and who have served in editorial roles for prominent leadership journals. Further, each book has been through many editions, providing multiple opportunities for revising and correcting mistakes in earlier editions and for adapting to continuing changes in the field; these three volumes have been through a total of nineteen editions, the earliest one in 1981 and the most recent in 2020. So, even scholars who distrust these definitions can, I hope, recognize that they provide a worthwhile starting point, in the expectation that this essay itself may then serve as a worthwhile starting point for better mapping out the conceptual foundations of the field.⁴

The three textbooks are these:

• The seventh edition of Leadership: Theory and Practice, by Peter Northouse (Northouse 2016). Citations will be abbreviated as (N, page number).

• The third edition of The Nature of Leadership, an integrated collection of essays edited by John Antonakis and David Day (Antonakis and Day 2018). I will refer to it as “Antonakis,” both for simplicity and since the key sections are virtually identical to the material attributed to Antonakis in previous editions without this co-editor. Citations will be abbreviated as (A, page number).

• The ninth edition of Leadership in Organizations by Gary Yukl and William Gardner (Yukl and Gardner, 2020). I will refer to it as “Yukl,” for the same reasons. Citations will be abbreviated as (Y, page number).

Here are their definitions:

• Northouse: “Leadership is a process whereby an individual influences a group of individuals to achieve a common goal” (N, 6).

• Antonakis: “Leadership is a formal or informal contextually rooted and goal-influencing process that occurs between a leader and a follower, groups of followers, or institutions” (A, 5).

• Yukl: “Leadership is the process of influencing others to understand and agree about what needs to be done and how to do it, and the process of facilitating individual and collective efforts to accomplish shared objectives” (Y, 25).

Constructing Criteria: What Makes a Good Definition?

The immediate question is a philosophical one. What are their criteria? Because these books are not intended to be philosophical, it is no surprise that the authors do not dwell upon this question. For the record, Northouse does not raise it; he moves directly from an entertaining overview of the “prolific stew” of definitions served up over the decades to his own, with the preface: “Despite the multitude of ways in which leadership has been conceptualized, the following components can be identified as central to the phenomenon…” (N, 6). Yukl nods towards the difficulty of the question of criteria when, in an earlier edition of his book, he remarks: “Like all constructs in social science, the definition of leadership is arbitrary and subjective” (Yukl 2010, 8). But then a basic pragmatism emerges, as he later writes, “…what matters is how useful the definition is for increasing our understanding of effective leadership” (Y, 40). Antonakis, for his part, rightly thinks that we get some help from ostensive definition, noting that “…leadership is often easy to identify in practice…” (A, 5).⁵ Knowing it when we see it is, indeed, often good enough. And Antonakis does frame a general principle: “…A leadership definition should be independent of contextual constraints and moral orientations so that its pure form, its defining conceptual bedrock, is identified” (A, 68). But, like the others, he moves straight to his definition: “Even in the absence of universal agreement, broad definitions…are required…. For the purposes of this book, we use the following definition…” (A, 5).

Those who are doubtful about defining leadership may have more general doubts about efforts to define anything. I begin, therefore, with a detailed methodological section, in order to ensure that the banisters are properly installed on the staircase. (Those who are sure-footed enough to ascend without banisters may now advance to the section entitled “the behavioral dimension”).

Concepts Versus Conceptions

My first proposal is that it is important to keep straight the difference between defining a concept and defining its conceptions. This particular pitfall helps explain the astonishing abundance of leadership definitions that every author discusses, and that has bewildered commentators for decades. There can be indefinitely many conceptions of any particular concept. That is to say, there can be indefinitely many ideas about how a concept might be actualized.

Take the concept of a clock.⁶ We might define it as an instrument designed for the purpose of telling time. There are countless conceptions of clocks—that is, countless ideas about how we might design an instrument that tells time—ranging from ancient water clocks to modern atomic clocks. These conceptions vary, among other things, according to their mechanisms, their power sources, and their display methods. If each of us defined the concept of a clock according to our favorite conception of a clock, we would quickly find ourselves throwing up our hands in frustration about how arbitrary and subjective the project of concept definition is. And this is part of the challenge facing any author who attempts to define leadership. There are many worthy ideas about how leadership might be actualized in practice, varying widely according to cultures, contexts, purposes, resources, and personalities. These are conceptions of how to go about it; to make any of them necessary to the concept of leadership itself would be, I suggest, to make the same mistake as saying that a clock, by definition, is a battery-powered timepiece. This is the perplexity that, I suspect, Yukl is wrestling with when he writes,

Whenever feasible, leadership research should be…relevant to a wide range of definitions, so that over time it will be possible to compare the utility of different conceptions and arrive at some consensus on the matter” (Y, 2).

My only suggestion to him is that it is easiest “to compare the utility of different conceptions” if we are working with a single definition which expresses what Antonakis calls the conceptual bedrock of leadership, and which would thereby include all perfectly good ways of leading, but which would exclude perfectly good ways of doing things that are not leading.

We shall later see, to provide one concrete example, that all of these authors spotlight a leadership style that is highly collaborative. This can be one way of leading. Collaborative leadership is certainly a legitimate conception of leadership. But my word of caution is that if the conception is pushed to the point of utter egalitarianism, it is now a conception of some concept—some potentially valuable concept—other than leadership. Role equality, and influence equality, have merit; but misunderstanding is sure to arise unless we identify them as something other than leadership. And the mistake takes root if we sow the seed of that conception into the very concept of leadership, thereby, as mentioned previously, propagating what I have called an anti-leadership bias.

Yukl might still contend that multiple definitions of a single concept are justified on the grounds that our definitions—and, thus our concepts—are sometimes revised over time in unexpected ways, as we gain new information about the world. To offer my own example, we no longer define “fish” in a way that would include “whales.” Our concepts of the two have changed. But fish and leadership, setting aside parables about fishers of men, are importantly different. Our fish concept is the concept of a natural kind, existing independently of human intentionality. Our leadership concept, however, is the concept of a social kind, fundamentally constituted by human intentionality. So I do not share the expectation that empirical experience might eventually tell us more about the concept of leadership itself; at the same time, I am optimistic that empirical experience will tell us more about conceptions of how we might better carry it out in practice.

In short, if we allow our permissiveness to extend beyond conceptions to concepts, it is likely to amplify what some scholars have called the big, loud, and sometimes disharmonious leadership chorus. A definition of the concept should help us to agree about which hymn sheet we are to sing from; once that is done, there will still be many beautiful ways of singing it.

Definitional Pragmatism

Leadership studies, being an applied discipline, brings to bear the insights of many academic disciplines upon the subject of leadership. In this way, it is like the study of education, journalism, politics, law, and medicine. Funders pay for the scholarship because, at least in principle, the results can be applied to the socially consequential practices of the teacher, the reporter, the politician, the lawyer, and the physician. Indeed, in all of these areas, scholarly work plays a critical role in the professional development of the current or prospective practitioner. Whether or not the leadership scholar personally cares about the practical benefits—one could engage in such scholarship for the sake of acclaim, royalties, tenure, or sheer fascination with the subject matter—I take it to be a truism that it could not count as leadership studies if it did not bear upon the practice of leadership.

Although all three authors explicitly endorse this position,⁷ I belabor it in order to caution against the temptation to think that scholarship would benefit from severing definition from practice. Yukl flirts with this idea when he writes, “The term leadership is a word taken from the common vocabulary and incorporated into the technical vocabulary of a scientific discipline without being precisely redefined” (Y, 22). But, as Joanne Ciulla asks rhetorically in her contribution to the Antonakis volume, “Would it make sense to have an academic definition that did not agree with the way ordinary people understood the word?” (Ciulla 2018,  441).⁸ In short, an important criterion for our definition of leadership is that it be consistent with the socially consequential practice that has given rise to its study to begin with. And if it turns out—as I argue in this essay—that formal and informal leadership are different practices, then not only do they require different definitions, but the degree to which they are socially consequential bears upon their merit for scholarly attention. In this way, I support the spirit, if not the exact substance, of Yukl’s pragmatism.

Necessary and Sufficient Conditions

Furthermore, since one point of definition is to prevent everything and nothing from counting as leadership, then the definition needs as much clarity as possible about what is out—the necessary conditions–and what is in—the sufficient conditions.

Some scholars resist this. One objection is that it is hard. As one academic worries, “I can see finding necessary conditions for leadership, but sufficient is a tall order.”⁹ It is important to see that it is worth the effort. To say what is out without saying what is in, or vice versa, is a failure to delimit, and thus, quite literally, a failure to define. All of the authors tacitly understand that this should be their goal in offering their definitions, even if they do not state it as such. One simple example is in the inclusion of influence by all three as a condition. They clearly mean it to be a necessary condition; they mean to say that any behavior that does not influence is out. But at the same time, they recognize that it is not by itself a sufficient condition, that it takes more than mere influence to be in. So, each of them elaborates upon the nature of the influence, so as to specify what needs to be added to influence to make it sufficient to be counted as leadership. I am not methodologically innovating; I am bringing closer scrutiny to bear on the existing methodology. In arguing that their efforts to specify what is out and what is in may have fallen short, I am merely extending their project.

A second objection is that such an enterprise is essentialist, and that essentialism is mistaken. Essentialism, in this context, is the claim that there is some independent metaphysical reality that we seek to locate and identify.¹⁰ This objection could be worth grappling with if leadership were a natural kind, but, as noted, it is a social kind. It is humanly constructed, relying upon no independent metaphysical reality. It is, accordingly, a practical inquiry. We can, nevertheless, meaningfully talk about seeking to define the essence of leadership, but only in the same way that we can meaningfully talk about its conceptual bedrock. These metaphors keep the language lively, but no prior training in mineralogy or ancient ontology is required.

A third objection is that overly precise necessary and sufficient conditions could threaten to distort the vague and elastic way that language works in the rough and tumble of everyday life. It is, indeed, usually inadvisable to introduce laboratory precision where it does not already exist.¹¹ To the extent, however, that leadership is to be defined in terms of things like organizations, goals, influence, responsibility, members, and followers, we may find that the definition can be no more precise than those constituent elements. Aristotle’s dictum is as apt today as when it was recorded in the Nicomachean Ethics: “We must only seek in each class of things such precision as accords with the subject matter” (Aristotle 1954, 11).

But this need not be a drawback, for two reasons. First, the definition can nevertheless assist us in advancing the conversation and eliminating confusion. If we agree to define “bachelor” with the two necessary and jointly sufficient conditions “unmarried man,” we may still find cases in which the notions of being married and of being a man are too imprecise to settle the matter. But we would at least know that we were ruling out of the conversation those who are bachelors only by virtue of having earned their B.A. And, second, scholars are always free to experiment with more precise versions of the constituent elements in suitable contexts. A government program designed to address poverty need not be paralyzed by the imprecision of the concept of poverty; it is free to announce, “for present purposes, we will take ‘poverty’ to mean ‘having an income of less than $17,609 for a household of one.’ This, appropriately enough, is called a precising definition.¹² It can be a useful scholarly project to explore whether increasing precision in specific ways makes any practical difference. But, to build on the earlier section on concepts versus conceptions, these would be conceptions of leadership that are already under the definitional umbrella of the broader and less precise concept.

A fourth objection points back to the mid-twentieth-century attacks upon logical positivism which targeted the analytic-synthetic distinction. W.V.O. Quine, in particular, cast doubt upon the possibility of such a thing as exact synonyms, except in the case of stipulation. This, of course, could cast doubt upon the current project, since the necessary and sufficient conditions of a definiens would need to be synonymous with the definiendum. Even Quine, however, does not argue that we should eliminate talk of synonymy from our vocabulary (Quine 1951, 20). As long as we recognize that perfect synonymy may be an illusory ideal, identifying rough and ready interchangeability of expressions can continue to serve many practical purposes. And what we have before us, I have argued, is au fond a practical purpose.

Descriptive or Prescriptive?

There are two broad sorts of definitions between which we must choose: descriptive and prescriptive. The descriptive variety, in the form of lexical, or dictionary, definitions, is the more familiar of the two. One very good dictionary, for example, provides 56 related ways that English speakers use “lead,” beginning with “to go before,” and then defines “leader” as someone who does any of them, and “leadership” as the doing of any of them (Flexnor 1987). As a dictionary definition, it provides a backwards-looking account of how the word has been used, and it can thereby be true or false.¹³ Another descriptive element of meaning, also backwards-looking, truth-value-bearing, and found in the dictionary, is etymology. Etymology can indeed spark insights into a word’s definition, but can never determine it. It is suggestive to learn that “lead” derives from an Old English word meaning “to go” or “to travel,” but the word has traveled a long way, and taken on much additional baggage, in the ensuing millennium. Prescriptive definitions—which include precising definitions—look to the future, and although they may have utility, they cannot be true or false. Being stipulative, they represent a proposal for agreement on a term’s meaning for the sake of a particular conversation. If we agree to the stipulation that our conference is about financial institutions, and you present a paper about riversides, you have, in effect, broken the bank.¹⁴

Given the link to leadership practice, we do need a definition of leadership that is descriptive, that is, one that truly reflects a concept recognizably deployed by practitioners. But it must go beyond a mere dictionary definition, since dictionaries do not aim to provide necessary and sufficient conditions.¹⁵ At the same time, there must be a prescriptive element. When a term can intend more than one concept—when, that is, there are multiple true definitions with different necessary and sufficient conditions—in order to avoid confusion we need to stipulate the one that we propose to talk about. Your candidate may have leadership in the polls, but at the same time may have no leadership qualities. We can provide different true descriptive definitions for each use of the term; but we still need to take the additional step of prescribing which of the two we are now going to talk about.

Ludwig Wittgenstein’s influential account of definition is illustrative (Wittgenstein 1953, 66). His idea is that for some terms, their various uses have in common nothing more than a family resemblance; a few family members have square jaws, one or two of those plus a few others have close-set eyes, and so forth. When you explicate the meaning of different uses of such a term, you find in each case a different and overlapping cluster of properties with no single common element. His famous example is “game,” but “leadership” would serve as well. If we sought a definition that embraced every member of the leadership family, we would end up with a long and unhelpful list of disjuncts. “To lead” means: “to hold the highest rank” (as a candidate in the polls), or “to be on the way” (as roads to Rome, or doors to the basement), or “to begin the story” (as a newspaper reporter), or “to tempt into evildoing” (as we ask the Lord not to do, in the Lord’s Prayer), or “to receive the most screen time” (as a film actor), etc. They do have family resemblances, but we are not here engaged in a family business. So, instead, we try to describe accurately our one or two favorite family members, and we stipulate that, for this conversation, the others are to be neglected.

To recap, the standards which I propose are these: first, the definition needs to account for the concept in question at its most general level, without either requiring or ruling out any particular conception of how to go about realizing the concept; second, it needs to accurately describe the concept as exemplified by the behavior of its practitioners; third, if there is more than one such bedrock concept—as I shall argue there is, in the case of formal and informal leadership—then the extent to which the concept is socially consequential is one important factor in prioritizing scholarly attention; and, fourth, insofar as it is possible, the definition needs to make clear what is out and what is in—that is, it should seek to offer necessary and sufficient conditions.

The Behavioral Dimension

My analysis, as indicated, is organized according to four key dimensions which, I suggest, can be helpful in examining any definition of leadership. These are drawn from reflection about what would need to be entailed by any concept of leadership worthy of rigorous applied study. They are:

• 1. The behavioral dimension. What does, or ought, the leader do, that makes it leadership as opposed to some other activity?

• 2. The asymmetrical dimension. How is the leader differentiated from others in the group?

• 3. The social dimension. What is it that the leader is leading?

• 4. The teleological dimension. What can we know about the direction in which the leader is leading?

The behavioral dimension pertains to what the leader distinctively does, or ought to do, that makes it leadership. I take it to be uncontroversial that leaders are leaders by virtue of something that they either do or ought to do, which thereby constitutes leading, or its rough synonym, leadership.¹⁶ Some authors have noted that there is, in addition, a personal trait of leadership, and all three textbooks explore this idea.¹⁷ But, since the trait would be the propensity or ability to act in ways that leaders act, or ought to act, we still find the doing of, or the propensity to do, or the ability to do certain things at the heart of leadership. So, the study of leadership needs to be able to account for the actions, behaviors, practices, performances, or accomplishments that make leadership what it is or ought to be (as opposed to, say, making it some other practice such as, say, scholarship, stewardship, or horsemanship).

It is interesting to discover, then, that all three definitions seem to be carefully worded so as to avoid commitment to anything at all about the leader’s behavior. Antonakis, for example, defines leadership a “goal-influencing process” which occurs among the relevant parties. But he does not say who does the goal-influencing; we assume it is the leader, but that is left open. It would be consistent with this definition if the only goal-influencing was originated by the followers, while the leader was inert.¹⁸ So, this could equally well be a definition of, say, “followership.” Likewise with Yukl: if leadership, to use his language, is a process that influences and facilitates shared accomplishments, what prevents us from defining, say, “teamwork” in precisely the same way? And when Northouse defines leadership as “a process whereby an individual influences…,” it is reasonable to ask just who this individual is, and what the leader’s connection is to the process. I recognize that all three authors have much that to say elsewhere that is helpful and true about what leaders typically do or ought to do (that is, they describe conceptions of leadership). But it is fair to ask that their definitions indicate what leaders distinctively do or ought to do that makes their behavior leadership, enabling us to see why their typical behaviors are or are not instances of the concept. We are not there yet if the definition of leadership could, without further revision, serve as a definition of activities like followership or teamwork. Perhaps we have been told what is in; but we have certainly not been told what is out.

This silence does not initially puzzle the reader, I suggest, because we all bring along our inchoate prior notions of leadership. When we think of leaders, we conjure a mental image of someone who is or ought to be taking initiative, moving things along, making things happen. So, they mention influence, they mention common goals, it sounds good. They have not actually said that it is the leader who wields the influence or expedites the process, but they have not ruled it out either, and so we fill in the gap. But we can also begin to see why this is going to be hard to fix. It is in part, as I have forecast, because informal and formal leadership are separate concepts, with different necessary and sufficient conditions. If we try to improve the definitions to be explicit about the behavior of the leader which makes it leadership, we are stymied.

Here is the problem. Suppose we explicitly state that leadership is the actual initiation of an influence process; that certainly gets us closer to an account of informal leadership. But it entirely misses the central element of formal leadership, namely, that it is a matter of doing whatever it is that ought to be done in the leadership role, which may or may not call for initiating an influence process. But what if we go in the other direction and define leadership as doing whatever ought to be done in the leadership role; we have now left out informal leadership since the informal leader, qua informal, does not have any such role, and thus there is nothing that ought to be done in the leadership role by that person.

Formal versus Informal Leadership

The three authors unmistakably intend to include both the formal and the informal in their definition of leadership. Antonakis says so in his very definition. Neither Yukl nor Northouse mentions the distinction in his definition proper, but each plainly intends to be covering both. Yukl, for example, says elsewhere: “Leadership may be exhibited both by formally selected leaders and by informal leaders” (Y, 22). And Northouse states: “The leadership approaches we discuss…apply equally to assigned leadership and emergent leadership” (N, 9). They, in short, are defining it as a single concept that can appear in two different settings. This is not superficially worrisome; it is as if we were defining “dining” and making clear that we intend to capture both dining in and dining out with the same definition.

Note that each author uses common synonyms for both notions; so, we must bear in mind the Quinean point that perfect synonymy may be nonexistent, even while at the same time we acknowledge the rough-and-ready interchangeability of expressions. In the place of “formal,” they sometimes use terms such as “appointed,” “assigned,” and “selected.” In the place of “informal,” the preferred alternative is “emergent” or, in some cases, “emerged.”¹⁹ Very little is said about the boundary between the two, probably because it is taken to be of little importance. Northouse elaborates upon it a bit, noting that formal, or, as he prefers to say, assigned, leadership is “based on occupying a position in an organization,” examples being “team leaders, plant managers, department heads, directors, and administrators….” But someone else can become the “real leader” if that person is perceived as “the most influential member” (N, 8). But now a boundary seems to have been crossed. Northouse is surely right in saying each of two things: first, formal leadership is based on occupying a role or position; and, second, informal leadership is based on being highly influential. But these are two entirely separable ideas (allowing that being highly influential can be one possible path towards assignment to a leadership role). We are witnesses here to a quiet collision of two concepts when Northouse says that the most influential member of a group just is the “real leader.” The real informal leader, who wields the most influence? Or the real formal leader, who has the most access to organizational resources?

For clarity, let us start with some proposed definitions:

• Formal leader: A person who has the responsibility and rights for moving the organization in the right direction.

• Formal leadership: Whatever the formal leader (see that definition) does that pertains to that person’s responsibility and rights for moving the organization in the right direction.

This definition can apply to organizations of any size, including nested ones, and can pick out situations in which there is no leader (nobody seems to have such responsibilities and rights) and in which there is shared leadership (they are shared.)²⁰ Formal leadership concerns the socially constructed position, role, or office of the leader. In this way it is like any other such social construction, be it professor, publicist, or plumber. The position is created and filled by whoever has the power to create and fill it. This power can vary widely—it could be, for example, the power of a hierarchical superior, a council of elders, the police, or popular acclaim. Any position, I propose, just is the rights and responsibilities of the office; so, call these official rights and responsibilities. Because of them, there are certain things that, as the formal leader I can do, owing to my official rights,²¹ and that I ought to do, owing to my responsibilities. This is what makes formal leadership a functional notion. This will be explored later in more detail, but suffice it to say for now that the formal leader has a particular function in the organization—the leadership function. The official rights and responsibilities pertain to that function. And—relevant to this section of the essay—the behavior of the formal leader is whatever the leader does that pertains to that function, whether positive or negative.

By the expression leadership behavior, we often mean good leadership behavior, such as when we say that someone showed, or failed to show, leadership.²² But there is a broader sense in which behavior that constitutes poor leadership is itself leadership behavior; nobody misunderstands us when we say that Caligula’s leadership was catastrophic. And what about influence? Influence can be just as significant a factor in poor leadership as in good leadership; a formal leader can be a bad influence. And much formal leadership behavior has nothing to do with directly influencing followers—which should not be surprising, given that the concept of moving an organization forward only entails influence at the organizational level, and entails nothing about interpersonal influence. It would be an improbability, but not a contradiction, to say that the formal leader, despite being a bad personal influence, was all things considered a good organizational leader.

Let us now define the informal concept:

• Informal leader: A person who influences others to become, or remain, that person’s followers.

• Informal leadership: Whatever the informal leader (see that definition) does that influences others to become or remain the leader’s followers.

The only behavior of mine that is relevant here is that which influences people to become, or remain, my followers. It could be behavior that inspires trust, fear, mimicry, or simply the belief that I can serve their self-interest. The intentions behind my behavior increase the probability of my having an influence, but that is as far as my intentions go. They are not sufficient; I can entice, cajole, and bribe all day long, but if nobody notices me, I am not a leader. Neither are they necessary; I can accidentally gain followers. So can fictional and dead people.

If I am an informal leader, but not at the same time a formal one—if, that is, I have followers but do not at the same time have any official leadership rights and responsibilities—then none of the things that the official leader ought to do, or can do, apply to me. If I am merely an informal leader, then the board of directors, for example, cannot fault me for failing to lead the organization in the right direction, since it is not something I ought to do. Nor will the board consider next year’s budget to be approved if my signature is the one on it, since approving it is not something I can do. As soon as the board (or whomever has the power) does hold me responsible for leadership, or does acknowledge my leadership rights, they have, de facto, made me the formal leader.

De Facto and De Jure Leadership

Furthermore, in accord with the principle of focusing upon the most consequential concepts, it is de facto formal leadership that matters the most. If formal leadership is to be equated with the behavior of someone whose leadership responsibility and rights are socially sustained by the powers that be, then mere de jure leadership is neither necessary nor sufficient (and, practically speaking, of only secondary interest). De jure features pertain to the way things appear to be, as represented by the org chart, the bylaws, the job descriptions, and the policy manual. De facto features pertain to the way things actually are, the way things actually get done in the real world, which is determined by where the responsibility and the rights—most specifically, the power—really lie. Appearance and reality may or may not be the same.

This distinction has nothing to do with the boundary between formal and informal leadership. Note that we can always ask, in a particular case of formal organizational leadership, whether we mean de jure or de facto leadership. That is to say, we can always ask whether someone is the apparent formal leader or the actual formal leader. This is not a question about who appears to have, and who really has, influence. It is a question about who appears to have, and who really has, the leadership responsibility and rights. It would make no sense to ask, in a case of informal leadership, whether it is de jure or de facto. It is always de facto. It can certainly occur within an organization, but is not a feature of organizational structure at all. Informal leadership is a certain sort of influence over an aggregated group of individuals who may or may not happen to also be nested within an organization.

So, to return to the Northouse remark about real leadership: it is one thing to be the person who has the most individual influence in the room, or the “real” informal leader; but it is quite another to be the person in whom official rights and responsibilities have been vested by the powers-that-be, or the “real” formal leader.²³ I must, then, respectfully disagree with Antonakis when he says that, with regard to the definition of leadership, “whether one has formal authority or not does not matter.” (A, 6) What I am suggesting is that having formal authority can sometimes be what matters the most. I can try to influence you to go to war. Or I can declare war. There is an unmistakable difference in the two. The uses and misuses of formal authority—that is, of official rights and responsibilities—can be of the utmost importance, both for leadership scholars to study and for practitioners to learn. Similarly, I have a concern when Northouse writes: “When a person is engaged in leadership, that person is a leader, whether leadership is assigned or emerged.” (N, 9) “Engagement in leadership” can be an entirely different activity under the two different descriptions.²⁴ The informal leader, qua informal, has no official leadership responsibilities and rights (and thus no official authority), and thus cannot exercise them; the informal leader, thus, is “engaged in leadership” only when influencing individuals to become or remain followers. The formal leader, however, is “engaged in leadership” when doing anything that pertains to official leadership rights and responsibilities, whether or not any individuals are being influenced. The two, as noted, often overlap; but often they do not.

More on the Behavioral Dimension

All three authors, as we have seen, have largely restricted their definitions to interactive or influence-related processes.²⁵ This invalidates the leadership significance of any behavior that is not necessarily interactive or related to individual influence. Take, for example, this remark from Northouse: “The key assertion of the functional perspective is that the leader is to do whatever is necessary to take care of the unmet needs of the team” (N, 380, italics mine). This is at the same time both absolutely correct for all formal leadership, yet also absolutely inconsistent with his definition of leadership. The leader, indeed, has the responsibility to do whatever is necessary for moving the group forward. That may well include influencing them as individuals to act in certain ways. But what if “the unmet needs of the team” are such that the leader must go and get more resources, or hire a consultant, or find new clients, or make a speech to the chamber of commerce, or testify before a Senate subcommittee? None of these counts as an internal, interactive, social influence process, and so they cannot count as leadership on any of these three definitions.

Think of the famous historical example of leadership set by Teddy Roosevelt as he valiantly rallied the Rough Riders up San Juan Hill. As their formal leader, what he needed to do for the group was to generate individual followers, and he did it brilliantly. So, that instance of excellent formal leadership incorporated informal leadership. Imagine, however, that as formal leader he also needed to arrange for reinforcements before the charge in order to hold the conquered hill, but that he negligently failed to do so? That behavior would have had nothing to do with influencing his followers; but it would clearly have also been leadership behavior—very poor leadership indeed.

In Jungle of Stone, William Carlsen tells the little-known story of John Lloyd Stephens, a writer, diplomat, and visionary of the mid-nineteenth century. I provide an extended example from Carlsen’s narrative, because it provides a paradigmatic example of formal leadership; yet it would not be identified as leadership behavior based on these definitions, since the initiation of internal interactive processes is not essential to any of the behaviors described. Stephens, we are told, became an executive with the Panama Railroad Company. Its mission, prior to the building of today’s Panama Canal, was to connect the Atlantic and Pacific Oceans via a system of railroads and canals through the jungles of Panama. The company’s large workforce in Panama was severely hampered by terrain, weather, disease, and a shortage of resources; Stevens left them in Panama and returned to New York. The passage below describes his leadership behavior:

When he arrived home [in New York}, Stephens was elected the company’s new president, a development that must have felt as much a burden as a distinction. After his…many trips across the isthmus [of Panama], he knew intimately the enormous obstacles that lay ahead. With the full weight of the company now officially upon him, he would not be able to return to the isthmus for another six months. During the interval, his life was consumed by work: arranging contracts for locomotives, ordering rails, creating strategies to attract workers…. Everything had to be bought or built in the United States…and then transported to Panama aboard steamers. Even the housing for the hundreds of workers now converging on the isthmus was prefabricated and sent down for assembly, along with thousands of pounds of food and medical supplies. Doctors were hired and a hospital was under construction on the site. Steam pile drivers, wooden pilings, railroad ties, tools—in short, everything necessary for the construction of a railroad—all had to be bought and transported to Panama at great expense. Stephens had now become, in effect, king of one of the greatest construction projects of his day, a monumental undertaking…. He had to retake the jungle and remake its landscape…. His health still compromised, the work never-ending, he had no choice but to somehow summon up the energy to keep the dream of the first intercontinental railroad in the Western Hemisphere alive and moving forward (Carlsen 2016, 425).

Note, first, that a merely informal leader—someone who merely exerted influence upon followers—could have engaged in none of these behaviors, since they required the exercising of rights (to sign contracts, create strategies, hire people, etc.); nor would a merely informal leader have been expected to do so, since they pertain to carrying out leadership responsibilities. Second, even though the behavior is mostly external to the organization, it is its pertinence to carrying out these essential tasks that makes it leadership. We can identify places where influence was likely to have been exerted, such as making deals and hiring staff; but the fact that these activities involved influence is not, I suggest, what makes them leadership. What makes them leadership, on my account, is that they were what Stephens needed to do in order to carry out his leadership responsibility to advance the organization. The organization’s members were thousands of miles away, and most of those in the organization were largely unaware of his activities, or even his existence. The fact that organization members are so often unaware of such activities, given that they are by definition external activities, may help explain why they seem to be excluded by these definitions and neglected in leadership studies.

The Asymmetrical Dimension

We now turn to my second broad concern, namely, the presence of an anti-leadership bias in leadership studies, which is demonstrated by what I call the asymmetrical dimension. This dimension of leadership, I contend, is conceptually indispensable. Leadership is enormously interesting, so interesting that it generates textbooks, monographs, journals, articles, self-help books, conferences, centers, degree programs, courses, webinars, op-ed pieces, and gurus. It does this, as I have noted, because it is socially consequential. And it is socially consequential for one of two distinct reasons: either the leader has personal influence; or the leader has official responsibilities and rights. In each case, the asymmetry is different. In the case of informal leadership, there is necessarily an influence asymmetry between the leader and those influenced. If everybody is equally influencing everybody else in some respect, then it might be an instance of utopia, groupthink, or mass hysteria; but it is not informal leadership. In the case of formal leadership, there is necessarily an asymmetry with respect to responsibilities and rights. If everyone has the same responsibilities and rights, then it might be an instance of a benign Rousseauian state of nature, or perhaps of a bellicose Hobbesian state of nature; but it is not formal leadership. In the prevailing definitions these necessary asymmetries disappear; leadership is leveled.

Leadership asymmetry does appear elsewhere in these volumes, as it does throughout the literature, but in their definitions it is disregarded as inessential and, apparently, undesirable. This is most prominent in Northouse, whose concept of leadership is something that is “available to everyone” (N, 6), and who insists: “Leaders are not above or better than followers” (N, 7). That leaders are not better than others in the group is an important point; but they are “above” followers in one of two particular respects. They are above them either with respect to influence, or with respect to position. This is why the early history of leadership studies—and here I refer especially to Plato’s Republic—is preoccupied with the problem of how to ensure that leaders do not leverage this advantage in order to gain superiority in other respects.

Yukl, for his part, notes with apparent regret that “most leadership literature over the past half-century has focused on leaders,” before launching into his chapter on the importance of followers (Y, 275). And Antonakis offers what could be taken as an apology for his definition of leadership science, which he fears may be seen to be “rather leader-centric,” offering reassurances that followers are important too (A, 5). I do not find fault with these sentiments; they express what may or may not be defensible preferences for how an organization should get things done. I only point out that a definition of leadership must be leader-centric, or it is a definition of something else. This is perfectly consistent with there being conceptions of leadership—ideas of how a leader might go about practicing leadership—that have elements that are not leader-centric, just as there are conceptions of film directing that have elements that are not director-centric. In the case of formal leadership, for example, I could carry out my leadership responsibilities by promoting an egalitarian culture; but as soon as I go so far as to arrange for everyone to have the same responsibilities that I used to have, then I am no longer the formal leader, nor is anyone else.

Process

There is one thing that obscures not only leadership behavior, but also asymmetry, in each of these definitions: they express it strictly as a process. This is understandable. As Joseph Rost documents in his landmark 1991 study, the language of process has been used for leadership since at least the work of Emory Bogardus in 1934 (Bogardus 1934). Rost himself refers to leadership uncritically as a process, and the usage has been trending upwards for decades (Rost 1991).²⁶ Yet it is unclear what sort of leadership behavior the language of process is designed to pick out. It cannot be, for example, the action carried out by the transitive verb “to process.” That would mean that leadership is processing something, in the sense that you might process my argument, or you might process a piece of wool—perhaps not entirely different processes. If leadership is not processing something, then the term “process” must be intended as shorthand for engaging in a process, as with the process of baking cookies or running for re-election, or initiating a process, as with the process of filing for divorce or contesting a credit card charge.

So, let us test out this interpretation by looking at what each author says elsewhere about leadership as a process, and emend the language of each definition accordingly in terms of engaging in or initiating a process. Northouse, for starters, says that leadership is not about the leader, but is a “transactional event” between leader and followers. He further clarifies:

Process implies that a leader affects and is affected by followers. It emphasizes that leadership is not a linear, one-way event, but rather an interactive event. When leadership is defined in this manner, it becomes available to everyone. It is not restricted to the formally designated leader in a group (N, 6; see also N, 7).

So, for Northouse, it would make sense to say that the leader is engaging in a process, but only in exactly the same sense that everyone in the group engages in the process. And, since the process is non-linear and “available to everyone,” the leader is no more an initiator of the process than is everyone else. It is a process in which everyone can be the leader–or maybe no one. “Process” serves as a cloak, not a clarifier. And what it seems to cloak, even if inadvertently, is the preference for a definition of leadership that allows, as a substitute for leadership, interactive group processes in which all are equals.

How about Yukl? He explains, “Researchers who view leadership as a shared, diffuse, process,” that is, researchers who define it Yukl’s way,

…are likely to pay more attention to the complex influence processes that occur among members, the context and conditions that determine when and how they occur, the processes involved in the emergence of informal leaders, and the consequences for the group or organization (Y, 27).

In this account of leadership as a process, mutual influence among members rules supreme. Informal leadership is the only sort of leadership that is mentioned, but we cannot be sure what even that means, since “leadership” remains undefined until we understand why “process” is in the definition to begin with. Apparently, again, every member who engages in or initiates the process of influencing²⁷ is a leader, and every member is a candidate to do so. Yukl does say a few pages later, “In this book, leadership is treated as both a specialized role and a social influence process” (Y, 26). But given that he has already defined it exclusively as a process, he has no conceptual latitude to go both ways. This puts him in a difficult intellectual position, as we can see in these two consecutive sentences:

• The focus [of the textbook] is clearly on the process, not the person, and the two are not assumed to be equivalent.

• The terms leader, manager, and boss are used interchangeably in this book to indicate people who occupy positions in which they are expected to perform the leadership role.

So, to track his logic, the leader occupies a position in which the leader is expected to perform the leadership role. Except, that role is a process which is not equivalent to anything done by the one in the leadership role. The definitional engine locks up at this point, only having taken us far enough to see that Yukl, every bit as much as Northouse, has opted for a definition that allows for equality.

Antonakis concurs; his reason for couching his definition in terms of process, he says, is to incorporate “the interaction process between leader(s) and follower(s)” (A, 6). As in the other cases, this tells us nothing distinctive about the leader. The leader engages in interaction, but so does the follower. There is no reference to whether the leader initiates it, or should initiate it. So, on this account as well, we remain unenlightened about what distinguishes a leader from any other member. His co-editor for the current edition of the textbook, David Day, offers further clarification: “Leadership…is a dynamic process in which the onus of leadership can change quickly from one person (formal or informal leader) to a dyad or groups. It is not that there are leaders and followers and these are distinct people. It is that anyone can be a leader even when not the leader in an in-the-moment give-and-take among people engaged in shared work.”²⁸ In Day’s case, it seems, we have reached the bedrock conception of leadership when we recognize that leadership is a process in which “it is not that there are leaders and followers,” but rather “anyone can be the leader even when not the leader.” Day’s concept of leadership thus evaporates into the mist of equality.

The Anarchy Option

This, then, leads us to the third category into which I propose we sort recent leadership research. We have explored two categories: formal and informal leadership. This third category is not leadership at all, but egalitarianism—or, what most social philosophers would call anarchism. It represents a rich literature of opposition to social hierarchy, whether governmental, ecclesiastical, or corporate. The tradition is consistent with many of the process-oriented theories described in each textbook. It is driven by the noble view that people are better able both to achieve their common goals and to realize their full humanity if allowed to coordinate their activities independently, without the superimposition of leadership. This has never, however, been described by social philosophers as the view that “leadership is available to everyone” or that “anyone can be a leader even when not the leader.” It is called “anarchy” for a reason. In the Greek, an means “without” and archos can mean “leader.” There is no leader. “Pan-leadership” is no leadership. Any line of research that argues that influence, responsibility, and power is better diffused throughout an organization rather than concentrated at the top can count itself as a part of a proud and respectable intellectual tradition. But the research topic, then, is social dynamics, or organizational teamwork, not leadership, except insofar as anti-leadership research would be included in leadership studies as a critical foil.²⁹ Definitions of leadership that are so unruly as to recognize anarchism as anything other than an antonym are a long way down the path to permitting everything and nothing as leadership.

Day seems to consider this entire philosophical tradition to be irrelevant. He defends his pan-leadership approach, writing: “That is not ‘anarchy,’ but how work actually gets accomplished in organizations.”³⁰ But anarchy, in this well-known non-pejorative sense, could hardly be better captured than by Day’s own words, as the case in which “anybody can be the leader even when not the leader.” Pan-leadership anarchy can get organizational results—for example, when Sally, Buddy, and Rob are working as television gag writers in the Dick Van Dyke Show. But organizational work also gets accomplished when people have different roles (including leadership roles), understand their roles, and perform them well. The point is not about efficacy. The point is that egalitarianism, or anarchism, is not a conception of leadership.

There are, on the one hand, plenty of cases in our disorderly lives in which it is unclear whether we are looking at formal leadership, informal leadership, egalitarian teamwork, or some organically evolving admixture of all three. On the other hand, all three could be present in an entirely orderly way. Imagine that I am a leader who believes that role equality throughout the organization is essential to our success. For that reason, I make use of my personal qualities to individually influence people in order to win them over to an egalitarian style. I would, therefore, be making use of informal leadership (influencing followers) in order to carry out my leadership responsibilities (doing what is necessary to move the organization forward), the outcome being an egalitarian culture. These cases are as worthy of study as any others.

In sum, asymmetry is an essential element of the concept of leadership, whether it be informal leadership’s asymmetry of influence, or formal leadership’s asymmetry of responsibilities and rights. By defining leadership strictly in terms of process, the opportunity is lost to differentiate between leader and non-leader. The resulting definitions are consistent with philosophical anarchism, in which there is no leadership. My thesis is that we can benefit by clarifying not only the distinction between formal and informal leadership, but also the distinction between these two and a third category, namely, egalitarian teamwork. All three can be organizationally significant, and nothing prevents them from overlapping. Insofar as it is possible, however, it can be helpful to grasp their conceptual differences.

The Social Dimension

We now move to the two remaining dimensions of leadership, the social and the teleological. We immediately face a hurdle; it is hard to say what the leader does with respect to either of these dimensions when the textbook definitions, as noted, are not explicit about what the leader does with respect to anything. Readers of these volumes seem not to have noticed this omission, bringing to their reading experience their own paradigm cases of leaders, and interpreting the definitions in a way that fits those assumptions. I shall, therefore, as such a reader, offer a paraphrase of each definition which retains the key ideas, while capturing what typical readers would likely take each author to mean. I drop, for simplicity of exposition, language that seems not to do definitional work, as well as language that I have already attempted to show to be untenable (in particular, the language of symmetry). These are my working paraphrases:

• Northouse: Leadership is influencing a group of individuals to achieve a common goal.

• Antonakis: Leadership is influencing the goal of a follower, groups of followers, or institutions.

• Yukl: Leadership is influencing others to agree about what needs to be done and how to do it, and facilitating individual and collective efforts to accomplish shared objectives.

The authors have every right to object that this is not what they mean to say.³¹ My response, I hope a fair one, is that this is what the reasonable and educated reader is likely to interpret each to mean, and it is the most charitable way to interpret each if we seek a definition that actually specifies a leadership behavior and differentiates it from non-leadership behavior.

I turn, then, to the social dimension. What both formal and informal leaders are leading are groups of other persons (although I shall argue that this means something different in each case). This may seem obvious, but there are concepts of leadership that are not necessarily social. The first of the 56 dictionary definitions, “to go before,” only requires a certain asymmetry, and does not include any behavioral, social, or teleological dimension. This would apply to Mt. Everest’s leadership as the tallest mountain on earth; but that is clearly not high on our list. Furthermore, as Northouse aptly notes, it is a different, non-social, concept when leadership gurus “teach people to lead themselves” (N, 6). These perfectly good uses of the term “leadership,” lacking a social element, are among the members of the family, to recall Wittgenstein, that fall outside the proper domain of leadership studies.

All three of the paraphrases identify leadership with the influencing of individuals or groups of individuals,³² and they identify these individuals as followers.³³ This seems to accurately specify the social element of informal leadership. But, I propose, the formal leader, qua formal, is leading something different: the group, or organization, as an entity. This does not rule out leadership behavior which influences individuals, or creates followers, within the formal group, as in the earlier Teddy Roosevelt example. The concept of formal leadership, however, is not defined by that behavior. What makes it formal leadership is the socially created obligation to influence, if anything, the direction of the organization itself. This may or may not include influencing, or making followers of, individual members. In the remainder of this section on the social dimension of leadership, then, I will argue two things: first, that formal leaders do not necessarily lead followers, and, second, that they do necessarily lead organizations.

Formal Leadership and Followers

I shall develop this argument at some length, since I recognize that it goes against the deeply etched grain of much of the leadership literature. Let me first state the case in its most general form. If you are the formal leader, any uncontextualized instance of your cultivating followers within your organization—for example, “Spend an hour mentoring David”—could represent either terrific or terrible leadership. You are a position holder, as noted, and your position carries leadership responsibilities. In order to carry them out, you must deal with constantly evolving internal and external conditions, with limited hours in the day, limited resources at your disposal, and unlimited choices about how to spend those hours and resources. Every hour that you spend influencing me as a potential internal follower carries with it an opportunity cost. It may be better leadership on your part—that is, a better way of doing what you ought to do to advance the organization—if you send me to training, replace me with someone smarter, cultivate a different internal group, spend the hour on brand-building, reread the Effective Executive by Peter Drucker, or take a nap. Formal leadership is holistic, since, again, it is about doing what I ought to do to advance the organization, and whether any one element of it counts as good leadership can only be evaluated relative to the whole.³⁴

In order to lay some groundwork for this point, let us establish a conceptual distinction between a member of a group and a follower of the group’s leader. Formal groups often exist, and thus have members, without having leaders; this entails that I can be a member of a group without being the follower of any leader. Followers, in addition, are understood in the leadership literature to be individuals who are influenced in some relevant respect by the leader. If I am in a leaderless group which then acquires a formal leader, I may not even know that it has happened, or I may otherwise be uninfluenced or utterly unaffected by the leader. So, there is no inconsistency in saying that I am a group member, but that I am not at the same time a follower of its formal leader. The concept of a member is group-relative; the concept of a follower is leader-relative.

My larger claim, then, is that if I have formal organizational leadership responsibility, then leadership, even good leadership, on my part is entirely consistent with most or even all of the members not being my followers. I propose that, broadly speaking, there are at least two types of internal relationships that a member can have to an organization. One is as a mere member. As a mere member, I do not necessarily have any official responsibilities for contributing to the organization’s success, but I do have official rights, or benefits, that are my chief reasons for maintaining membership. This, for example, is the relationship that I have to my college alumni association, the Auto Club, and the American Philosophical Association. The other relationship is as position-holder, which includes employees, conscripts, and volunteers. As a position-holder, as noted, I have official responsibilities for contributing towards the success of the organization; also, with a little luck, I have official rights to support the execution of my responsibilities. This, for example, is the relationship that I have to my employer, or to the boards on which I serve. I could be both a member and a position-holder in an organization, but that need not be so. I could, for example, be a member but not a position-holder in a group by virtue of being an American citizen who does not work for the government. I could, alternatively, be a position-holder but not a member of a group by, say, working as the groundskeeper at a country club where I do not hold a membership.

Mere Members need not be Followers

Consider, now, mere members, whose membership may be based entirely upon gaining the benefits offered by the organization. Why must we judge them to be followers of the leader? That is to say, why must we require that they are necessarily influenced in a relevant respect by the leader? Many of us would deny that we are, or ought to be, followers of our mayors, governors, or presidents—especially if we did not vote for them, or if we oppose their policies. But this does not change the social fact that they are the leaders of the organizations in which we are members. We rarely would see ourselves to be followers of whoever formally leads our health club, our HMO, or our homeowner’s association–indeed, we often do not know or care who they are. We do not care to be personally influenced by them. What we usually care about is that they do a good job of delivering member benefits; assuming that the benefits are our reason for membership, then that is part of what good leadership amounts to in this context. If such a leader recruits mere members to become personal followers, that could be reason for suspicion. If I am a mere member of a religious congregation and you, as the leader, try to influence me to follow you personally rather than the principles of our faith, I would be well-advised to avoid any Kool-Aid you might mix. If you, as my nation’s president, exhort me to follow you personally rather than the constitution, I would be well-advised to promptly unfollow you on Twitter. Emergencies can be a different story; the Roman constitution, for example, allowed for the six-month appointment of a dictator in a time of crisis. But under normal leadership conditions,³⁵ the notion of being a mere member is conceptually orthogonal to that of being a follower. And if an organization has nothing but a leader and mere members, then that leader—even if a good one—may, in principle, have no followers.

Position-Holders need not be Followers

What about the members who are also position-holders? This would seem superficially to be different. The leader and the other position-holders are, or should be, working in tandem towards the same organizational goal, with the leader, well, leading the way. But there are vast numbers of position-holders who would be puzzled, if not insulted, if told that they were, or should be, followers of the organizational leader. Just to be sure that our concept of following does not begin to drift, what this would mean—using the paraphrased definitions of the three textbooks—is that the position holders would be puzzled, if not insulted, if told not only that they were, or should be, individually influenced by the organizational leader, but also that the leader’s influence upon them as position-holders defines what it means for that person to be the leader. Consider, say, janitors, college professors, train conductors, partners in a law firm, bartenders, and elephant keepers. Such position-holders may value the leader highly, but only for leadership purposes such as hiring qualified people, properly equipping them, and getting out of their way. If the leaders do not individually influence each of the position-holders, this does not alter the social fact that they are the leaders of the organizations. Casey Stengel, the Hall of Fame baseball manager, was not entirely joking when he reportedly responded to a question about the secret of his success: “I try to keep the players who hate me away from the ones who haven’t yet made up their minds.” If an organization has nothing but a leader and self-sufficient position holders, then that leader—even if a good one—may have no followers. This is a good way to make sense of the much admired Taoist leader whose people, we read in Lao-tzu, “are hardly aware the leader exists. …When the work is done, the people say, ‘Amazing: we did it all by ourselves!’” (Lao-tzu 1988, 17).

I have already made a case against any effort to rescue the prevailing definitions of leadership by simply redefining “follower” as “member.” A less heroic rescue, I imagine, might seek to stretch our normal notion of following to encompass any deference, or even any acknowledgement of the need for deference, to the leader. But I suggest that this too would miss the point; such watered down guidance would seem no more compelling as an example of leadership than would the directions provided by a traffic officer who is filling in for a faulty traffic signal. Organizational leadership does matter for self-sufficient position-holders, but what matters about the behavior of the leader, what makes it good leadership, is not always that they are following the leader or even receiving any sort of guidance from the leader. It is often the other things that the leader brings to leadership; but these actions can no longer be considered to be leadership when it is defined, problematically, solely in terms of influencing followers.

How about the position-holders who are not so self-sufficient? Here, at last, we do find organization members whose following is not unlike the following of an informal leader. But I propose there is still a distinction to be made. Good formal leadership, under normal leadership conditions, is attentive to the Weberian distinction between the position and the person (Weber 2015). Because the organization may well endure after you and I are gone, it is best served if we are all clear that you are not, precisely, following me when you follow my guidance—or, if we are in the military, when you are following my orders. It is undeniable, of course, that my personal qualities—whether they include expertise, vision, charisma, trustworthiness, or even cruelty—can make it more likely that you will follow my guidance, or my orders, or my example. But leadership that is understood only in terms of influencing followers has no conceptual apparatus for noting that the person-position distinction, so important for formal leadership, is nonexistent for informal leadership.

This point is further underlined by the exception Yukl provides for “members of an organization who completely reject the formal leader…” They are not followers, he says, but “rebels” or “insurgents” (Y, 27). He is, of course, right that they are members but not followers. But he would also, given that leadership is defined only as follower-directed behavior, need to hold that there is no leadership behavior that would pertain to them. Yet, with a nod to Abraham Lincoln, I observe that the way a formal leader behaves towards rebels may be the greatest test of leadership. Yes, good leadership may require attempts to convert rebels into followers. But the best leadership may involve nothing of the sort; a good leader might crush them, fire them, marginalize them by demotion or reassignment, allow the situation to play itself out, or capitulate and resign for the good of the organization. If the social dimension of leadership is defined by leader–follower influence relationships, then there is no way to insist that a theory must also incorporate leadership behavior that pertains to anyone who is not a follower—whether they be mere members, self-sufficient position holders, or, now, rebels. Even when there are followers, formal leadership is necessarily about far more than the followers.

One interesting objection to my argument could be fashioned out of Antonakis’s explanatory remark that “It is largely because of followers that leaders are legitimized” (A, 6). This general idea has some currency in the contemporary literature on social ontology, with John Searle serving as the chief spokesman for the view that socially constructed facts, such as positions, are created by the assent of all of the members of the group (Searle 2009). If this is true, then a case might be marshaled for the thesis that formal positions are themselves ultimately constituted by—and thus defined by—the leader–follower relationship. Just as my following you is straightforwardly what makes you a leader in the informal context, my assenting to you serves the same purpose, but in a subtler and more profound way, in the formal context. It constitutes the ultimate source of your official responsibilities and rights as formal leader.

There are two problems with this potential objection. First, even if the social ontology is correct, it does not support the conclusion that members are followers, only that they are assenters. But, second, if I am correct in my case about the social construction of positions and their affiliated official rights and responsibilities, then Searle’s social ontology cannot be correct. The assent that matters, I have proposed, is the assent of whoever has the power to create and fill the leadership position. That could be the group as a whole, which is what we often find most desirable–although maybe not in prisons or elementary schools. But, alternatively, it could be the owner, the board of directors, the shareholders, a minister of central planning, the military, a divinity, my own personal ability, or some blurry blend of the above. So, again, even though followers are essential to informal leadership, there is a compelling case against them as a necessary, defining, condition for formal leadership.

Formal Leadership and Organizations

If formal leaders are not leading followers, then what are they leading? They are, as I have noted, leading organizations. Antonakis explicitly includes institutions as one of his disjunctive options.³⁶ Yukl may be opening the door to the same option when he allows for both “individual and collective efforts,” especially since he repeatedly notes that his concern is “managerial leadership,” and his book is entitled Leadership and Organizations. And Northouse probably means for his “group of individuals” to embrace both formal and informal groups. Throughout his book he consistently talks about “the function of leadership” and the one in the “leadership position,” neither of which makes any sense in an informal setting. But none of them makes room for the case that leading a formal group is conceptually different from leading followers.

Allow me to develop this case more fully, relying upon ideas introduced earlier in this essay. Associations of persons may be usefully characterized along a conceptual continuum, with one extreme identified as mere groups and the other as robust organizations. Mere groups are simple aggregates, in which the group is nothing more or less than the sum of the persons in it. Mere groups can exist either inside or outside of robust organizations. They are the domain of the informal leader. These aggregates may be small, such as a few children in a schoolyard; or they may be large, such as a stadium full of sports fans, or, perhaps, even all of society. But as we move across the continuum, aggregates take on socially constructed organizational features. If it is determined that one person will play a specific ongoing role, like collecting the trash, the group is now at least a proto-organization. If that ongoing role is leadership, then what may have been informal leadership might now morph into formal leadership; for what may have been personal influence upon an aggregate of individuals could now morph into responsibility for an ongoing unit.

Organizations arise at a point on the continuum where the group is able to acquire resources; the resources may initially be nothing more than the human resources harnessed by assigned roles. But resources can expand to include things like real estate, reputations, and legal rights—things that are owned by the collective, but not by any of the individual members. And they can acquire goals. These may initially be nothing more than the agreed-upon goals of a critical mass of the aggregated members; but organizational goals can take on lives of their own. The organization can then use its resources to advance its goals (which goals, carefully selected, can themselves serve as resources). This socially created organizational apparatus creates a group identity that transcends member identities, that persists regardless of changes in membership, and that becomes more and more robust as we move across the continuum.³⁷ It can even itself be considered a person for certain moral and legal purposes.³⁸ Formal organizations, so understood, can be small or large. They can be free-standing or nested deep within other organizations. And if a group is a formal organization—if it is able to have resources, goals, and an identity which do not hinge on the resources, goals, and identities of its members—then it is also able to have formal leadership.

So, formal leadership involves more than leading the group’s constituent persons. It necessarily pertains to social constructions that exist within organizations but that do not exist within an informal group. These include official responsibilities, official rights (including powers), resources distinct from member-owned resources, and goals distinct from member-espoused goals. None of these exists, or can exist, in an informal group, qua informal, and, thus, none of them is the concern of the informal leader. This is why we need not see it as more than a superficial paradox that, in principle, someone can exercise leadership, even good leadership, of an organization even while no one in the organization is the leader’s follower. Leadership responsibility for an enduring organization is one thing, while personal influence over a transitory aggregation of persons is quite another. If anyone insists that leadership, formal or informal, must include influence, my proposal is this: the only defining responsibility of a formal leader which has to do with influence is the influencing of the direction of the organization itself, and not necessarily its members.

The Teleological Dimension

I now turn to the teleological, or directional, dimension of leadership. My being your group’s leader entails something about the direction in which I am leading the group, or ought to be leading the group. This dimension is what makes it possible to talk about good or poor leadership. More specifically, it provides us with a way of framing two teleological questions that bear upon the assessment of leadership performance. To what extent is my behavior directed towards the right end? And to what extent does it deploy effective means in the service of that end? This essay, conceptual as it is, shall only be concerned with internal judgments of ends and means; by this I mean judgments relative to the group itself. There are also external judgments, or judgments from standpoints that are outside of the group. These may take into account the perspectives of a larger organization, of a stakeholder, of society in general, or, of course, of morality.³⁹ This is one of the philosophically richest topics in leadership studies, but cannot be developed here.

The teleology of informal leadership is straightforward: it consists in following the leader as a result of the leader’s influence.⁴⁰ The question about means, then, is likewise straightforward. The means is whatever behavior (intentional or unintentional, as noted) produces the influence that results in followers: in this single regard, then, the more effective the behavior, the better the leadership. As for the ends, there is no internal standard to assess them. They simply exist. The direction of the group, by definition, accords with whatever about me the members have seen worthy of following—maybe my ideas, my example, my projects, or my perceived ability to serve their self-interest. And it may be all of the above, since the ends of the group are nothing more than the aggregation of the ends of each individual member. So, strictly from the internal point of view of the followers, the ends are by definition taken to be worthy, right, and unassailable. If they are to be critiqued, it must be done by appeal to external standards—on the grounds that, for example, the group’s ends undermine larger organizations goals, are harmful to society, are sinful, or are unethical.

The teleology of formal leadership starts out in roughly the same way. We routinely judge formal leaders on the basis of their effectiveness, which pertains strictly to means. But, in the formal case, the means are organizationally (not individually) moored, which means that from the organizational point of view, there can be a right direction and a wrong direction. Effectiveness in the service of the wrong end—owing to corrupt interests, for example, or individual glory, or simply losing one’s organizational way—is indeed effectiveness, but it moves the organization in the wrong direction–effectively. This is why the concept of formal leadership must include the notion of the right direction; organizational responsibilities and rights that pertain to moving the organization are pointless if movement in any direction at all counts as good leadership. This issue does not arise with informal leadership, given that it is not the leadership of an organization, and thus not the leadership of a group which has an end that transcends its members’ individual ends.

Formal leadership, then, is doubly teleological. The first layer of teleology, which formal and informal leadership share, may be termed leader teleology. This simply means that what the leader does, or ought to do, essentially pertains to the direction of the group. But there is a second layer in formal leadership, which I shall label position teleology. Unlike leader teleology, position teleology obtains not just for the leader but for every position in an organization, regardless of the position’s rights and responsibilities. This layer of teleology has no bearing on informal leadership, which is by definition not a position.

Functional concepts, such as the concept of a position, can better be understood by way of contrast to results concepts. As an illustration, we can restart our clock example. A clock has a socially agreed upon function, which is to achieve a particular end, or telos: to tell time. So, if something is a clock, there is something that it ought to do: tell time. Compare a clock to, say, a tall rock that casts a shadow on the various objects around it in a way that happens to mark the time. Marking the time is a result of something about the rock, but it is not a socially created function. (Perhaps the rock could qualify as an informal clock.) If a gust of wind blew over the rock and you remarked that it was now a bad clock, I might plausibly respond that, no, it was not a clock at all. But if the same gust blew over our clock tower, my response would be different. I would say, yes, it is a bad, or broken, or defective clock; for it does not do what it ought to do. It does not perform its function. So, one way of evaluating the clock in the tower is functionally; this evaluation is, paradoxically, a matter of fact. It is not based upon whether I value either the clock or the practice of timekeeping, but upon how well the clock achieves its function.⁴¹ This evaluation is internal to the concept of the clock itself. The rock, however, has no timekeeping function. We can evaluate it according to external standards, but not according to its nonexistent function as a clock.

This can be applied straightforwardly to position teleology. Consider dogcatchers. We have, on the one hand, the functional concept. When you are elected county dogcatcher, your position is such that you have a responsibility to do things like catching stray dogs—and, if the county is well-run, you also have concomitant rights, like the use of a van and a net, to make the responsibility easier to carry out. Because of this socially created function, you are the dogcatcher regardless of whether you catch a dog; but whether you are a good one is another matter entirely. We also have, on the other hand, the results concept. If you are not the county dogcatcher, the only way to become a dogcatcher is to catch a dog. This second concept of dogcatching—call it informal dogcatching—is strictly a results concept. If it so happens that you have never done something that resulted in catching a dog, this does not mean that you are a poor dogcatcher. You are not a dogcatcher at all.

We can apply this analysis to other organizational positions, each of which has its own function. There are at least two ways in which I can be a marketer, a bookkeeper, a security guard, or a leader. There is the results way, whereby, whether or not I hold the position, a result of my behavior is that I market something, I keep books, I guard something, or I lead something. And there is the functional way, whereby I hold the socially constructed organizational position of marketer, bookkeeper, guard, or leader; in these cases, I ought to market, or bookkeep, guard, or lead. And in the case of the formal leader, its function is typically not defined by a marketing telos, or a bookkeeping telos, or a security telos; its function is defined by a teleological telos, namely, moving the organization in the right direction. The specific position itself imbeds leader teleology in position teleology. And, thus, formal leadership is doubly teleological.

The problem is not that our textbook authors deny this functional notion, but that they do not recognize that it is conceptually distinct, helping to explain why leadership studies seldom pay attention to this highly consequential dimension of leadership. All three of them talk about roles, positions, functions, and what position holders ought to do or must do. Antonakis goes so far as to say: “At its essence, leadership is functional….” And he is comfortable talking about non-relational behaviors that are a part of this function, such as, “leaders must monitor the external and internal environments…” (A, 7). The idea of functionality, and also the idea of non-relational behaviors, however, have already been ruled out of the concept of leadership, since the informal concept of influencing followers—which is strictly a results concept–has been baked in as a necessary condition.

They have, therefore, permitted no conceptual means for coherently talking about the organizational aspect of good or bad leadership. It is as though we have misdefined a clock, not as an instrument with the function of telling time, but as an entity that marks time. On that definition, there is no conceptual room for distinguishing between a good and a bad clock. If it is broken, it cannot be said to be a clock that is broken; it does not mark time, which logically entails that it is not a clock. If I am in a leadership position, and I am not doing what I ought to do to advance the organization, then according to these non-functional definitions I am not the leader. But if I am the one with the responsibility and rights for moving the organization in the right direction, I am indeed the leader. You might say that I am “no leader”, but that would not be literally true. It would be literally true only if I both had no followers, and thus was no informal leader, and also had no organizational responsibilities, and thus was no formal leader. In this case, rather, to say that I am “no leader” is shorthand for saying that I am a poor leader, or, like the toppled clock, that my leadership behavior is bad, broken, or defective. An exploration of the teleological dimension, then, continues to underscore the problems that are produced by assuming that informal and formal leadership are conceptually the same animal, one of them in the wild and the other in captivity.

Conclusion

I have argued in this essay that, taking these three textbooks as a proxy for the field of leadership studies, the current understanding of the concept of leadership can be enriched by a definition that is accurate in four respects: what leaders distinctively do or ought to do that makes them leaders; the asymmetry between leaders and non-leaders; what it is that is being led; and the nature of the directionality. My inquiry has pointed in the direction of three distinct concepts that are sometimes commingled: the non-leadership concept of egalitarian teamwork, plus the two separate concepts of formal and informal leadership, each with different necessary and sufficient conditions, and thus with different definitions.

First, and perhaps most subversive, then, is the concept of egalitarian teamwork. Leaders certainly may choose to lead in such a way as to cultivate a culture of egalitarian teamwork. But the equality that results would exist within the team, not between the team and the leader—unless the former leader is now an equal member of the team and, philosophically speaking, anarchy prevails. For the very concept of leadership requires asymmetry, either with respect to influence or with respect to rights and responsibilities. It is no insult to the notion of egalitarian teamwork to say that its elimination of both asymmetries would put it outside of the leadership domain. I call it perhaps the most subversive of the three concepts because its anti-leadership bias is surreptitiously admitted into the definition by virtue of the emphasis upon process, and by virtue of the reluctance to say who plays which role with respect to the process.

Second, and perhaps most conspicuous, is the concept of informal leadership. I define an informal leader as someone who influences others to become and remain that person’s followers. Informal leadership, then, is whatever is done by informal leaders which influences others to become or remain their followers. This concept is a results concept—that is, it applies only when the result of following occurs, regardless of the intentions of any would-be leader. And it can only be evaluated externally—from the point of view, for example, of morality, or law, or custom. But it cannot be evaluated as something that the leader, qua leader, ought to do by virtue of having the leadership function, given that the leadership is by definition not functional. I say that this is perhaps the most conspicuous of the three concepts because influencing individuals to become followers is the headline for every definition.

Third, and perhaps most consequential, is the concept of formal leadership. I define a formal leader as someone who has the socially created responsibility and rights for moving the organization in the right direction. Formal leadership, then, is whatever is done by formal leaders which pertains to their organizational responsibility and rights for moving the organization in the right direction.⁴² This may involve influencing individual members of the group, but doing so would not count, conceptually, as good leadership unless it played a role in moving the organization in the right direction. The concept is a functional concept—that is, there is something that the leader ought to do, given the social construction of the position. This means that, unlike informal leadership, there is an internal element to the evaluation of the end. Whatever I do that pertains to my responsibility is my leadership; but whether my leadership is good or bad is largely determined by how well I am executing that function. I say that this is perhaps the most consequential of the three concepts because, first of all, once I am in a leadership position, I retain whatever informal leadership ability I may have brought to it; if I was good at collecting followers beforehand, I do not lose that ability. But, second, I now have access not only to the resources of the organization, but also to official powers to deploy those resources. Even if I am merely the formal leader of a small organization nested in a larger one, my capacity for inflicting damage or generating benefit is magnified.

At the outset, I quoted Antonakis and Day’s lively likening of leadership research to assembling a jigsaw puzzle when the pieces have been drawn from several different puzzles. What I hope I have shown is that, while this is true, assembling it may be rendered less daunting by narrowing down the sources of the pile of pieces to these three distinctively different puzzles.