A Good Talk

The Story and Skill of Conversation


By Daniel Menaker

Formats and Prices




$11.99 CAD



  1. ebook $9.99 $11.99 CAD
  2. Trade Paperback $16.99 $22.99 CAD

This item is a preorder. Your payment method will be charged immediately, and the product is expected to ship on or around January 4, 2010. This date is subject to change due to shipping delays beyond our control.

A stylish, funny and surprising guide to the art of conversation, from one of New York’s foremost literary wits.

A GOOD TALK is an analysis of and guide to that most exclusively human of all activities– conversation.

Drawing on over forty years of experience in American letters, Menaker pinpoints the factors that drive and enliven every good conversation: the vagaries (and joys) of subtext; the deeper structure and meaning of conversational flow; the subliminal signals that guide our disclosures and confessions; and the countless other hurdles we must clear along the way. Moving beyond self-help musings and “how to” advice, he has created a stylish, funny, and surprising book: a celebration of “the most excusively human of all activities.”

In a time when conversation remains deeply important– for building relationships, for relaxing, even for figuring out who we are– and also increasingly imperiled (with Blackberries and texting increasingly in vogue), A GOOD TALK is a refreshing celebration of the subtle adventures of a good conversation.


Begin Reading

Table of Contents

Copyright Page



In this book I want to talk about the story and shapes and skills of conversation and also, ultimately, about conversations, no matter how transient they may be, as a kind of artifact—a human art of great importance produced by all people everywhere. But let's start at the beginning. We can be certain that this exclusively human activity called conversation didn't start out particularly shaped or aesthetical. No one can know how language began, but that hasn't stopped anyone from speculating about the matter.

I've always guessed that human language had to begin with grunts that developed into one or both kinds of pragmatically crucial "first" speech. One is naming or calling out of names—at the outset, of individual or groups of human beings, especially when they were out of visual range, as they might be in a hunting party; then of objects and animals and places and the weather. Something equivalent to our "John" must have been an early proper noun. Same with "Mary," "gazelle!" "hot!" "cave," "lion!" "not tonight," and "headache."

My other speculation about first speech is that it was orders or directions. Very close to naming. Let's say that the prehistoric equivalent of Mary had a daughter who would probably today be named Meghan. Let's say Mary was weaving a thatch of reeds for the roof of their hovel. Meghan, ten years old now and irritable about the fact that virtually nothing—from jump rope to television—had been invented, was mooning and sighing around the hovel with nothing to do. Mary needed more reeds, a stack of which her John had piled up outside the hovel, so she grabbed Meghan's arm and pointed toward the reeds and uttered something that eventually came to mean more or less "Fetch!" There followed a rash of people using this order all the time, until someone came up with the words "Oh, Mom!" and, probably later, "Get it yourself."

By the way, there is a word for the study of the origin of language: glottogony. It turns out that a lot of academics subscribe, roughly, to the naming/orders glottological theory that has always seemed to me to make sense. Biologists, evolutionary and otherwise, have made their own contributions to the understanding of the development of language by demonstrating that somewhere between a hundred thousand and fifty thousand years ago, the human larynx migrated, through mutation and natural selection, in such a way as to facilitate speech, perhaps as a result of Homo going erectus on the savannas and veldts. (A short-story submission I once read took place on what the author referred to as "the African svelte"—which, it now suddenly occurs to me, may be an accidental portmanteau of "savanna" and "veld.") It descended a little and assumed an L shape. The larynx, I mean.

There also appears to have been at around the same time—the time of hominids leaving the jungle for the plains—another crucial mutation in a certain gene, the FOXP2 gene (shown here so that you can recognize it whenever the need arises)—which led to the development of Broca's and Wernicke's areas (just south of the amygdala, off Exit 19). This important advance in the evolution of the human brain helped to facilitate speech. And it happened to all of them before the hominids left Africa, perhaps starting with Mitochondrial Eve, the mother who began all others. The mutation spread quickly because it was so helpful to hunting and gathering and was thus a natural selection for natural selection. That's why when the species spread out from Africa, it developed languages wherever it went.

One of the ways that scientists got onto this gene's role in language was the discovery not long ago that four generations of a modern family that had certain severe troubles in talking had a new, bad mutation at FOXP2. When you put all this and some other factors together—my guesswork, fossil evidence (of larynx bones), the deforestation of hominids, the FOXP2 gene, the need for foragers to go out on the svelte and come back with good directions to where the doomed antelope were playing that day, the use of tools, which required early man except the Jews and the Italians to cut down on the hand gestures and ramp up the larynx, and the real advantages of being able to communicate in the dark ("Faster!")—you get: language.

But language is not conversation, and conversation is what we are here to discuss, one of these days. It apparently could have been the case that language and speech preceded conversation by thousands of years. People, or something like them, may well have gone around just naming things for quite a while, expressing anger and ordering others around—you can still hear traces of this proto-language in New York City traffic jams—and conveying important information, such as "A saber-toothed tiger is coming" or "A little of this stuff that I herewith name 'garlic' improves the taste of antelope meat" or "The crushed and macerated leaves of that plant will help your acne." Those may be fine things to say, but they aren't conversation. Conversation started, I think, when, for example, to the first sentence a hearer responded, "I just want to tell you how much I've appreciated your help through the years, John," or to the second, "How did you discover that?" or to the third, "Great! The prom is this Saturday night."

That is, conversation can certainly contain and almost always does contain pragmatic information and pure expressions of emotion and so on, but for me, to be real conversation, it also has to include thoughts and ideas and reactions that are not simply reflexive and that have no immediate practical use. If it is all immediate usefulness, then we call it trigonometry class or a conference call or This Old House. (Not those useless but evidently still necessary primate rituals called business meetings, however—meetings are generally not much good for anything. Samuel Johnson once said that conversation is what's left when business is done.) So, then, let's take speech and language as givens, because they are. I mean, we all have them, even deaf people, whose sign language fits the definition of language in every way, unless you require audibility. I just thought that because we can't have conversations without them, it would be a good idea to give them a nod at the start. But now: How and why did the activity we call conversation begin?

One way to attempt to answer the question is to work backward from what conversations generally consist of to how they may have originated, and that is exactly what Professor Robin Dunbar (among others), of the University of Liverpool, has done in his book Grooming, Gossip, and the Evolution of Language. Dunbar postulates that conversation, especially gossip, began as a hands-free substitute for physical grooming, as hands were beginning to come in so… handy for other, newfangled purposes: tool making, sewing and reaping, cooking, cave painting. You could get real practical opposable-thumb things done and, at the same time, point out that the gigantic glacier seemed to be getting closer every day or that this new, very hot "fire" thing seemed extremely interesting. Now, when I say "gossip," you may not think of gossip as a real and respectable form of conversation, but you'd better. According to another social psychologist, Nicholas Emler, of the London School of Economics, more than 80 percent of our talk is about other specific and named individuals—in other words, gossip, or something very close to it. This leaves less than 20 percent for the meaning of life and cosmology and whether Cézanne's work anticipated scientific discoveries about the neurology of vision. (It did.) And many scholars have established that gossip gives us a way to examine ourselves and others with regard to in-bounds and out-of-bounds social behavior. So it is not only respectable and universal but also essential. And fun.

Here is a bit of gossip I shouldn't but will tell you. A colleague of mine once asked me as we were walking back to work after lunch if I would give him a recommendation for a job he was applying for at another company. I said that I was sort of under an obligation to act in the best interests of the company that employed both of us, and that he had already put me in a somewhat difficult position. He asked again, more insistently, as he was very unhappy in his current work situation. I told him that I really didn't think we should discuss the matter any further. He left the company shortly after that but not for the job he wanted. Later, I talked the situation over with a close friend, who ended up berating me for not helping the guy out. He said that I had been too rigid in following "the company line." I replied that at least I had kept the other guy's wish to leave a secret. "Shouldn't that count for something?" I asked. My friend said, "A little. Maybe." Now, do you see how many social norms these two instances of gossip involved? I sort of do, in fact. In any case, the received wisdom about gossip is that it serves a social-policing function.

Professor Dunbar also theorizes that the size of animals' social groups, especially primates' social groups, corresponds directly to the size of their neocortexes. In humans, our neocortex dictates that we can maintain social ties with about a hundred and fifty people per person. Chimps generally hang out with some fifty other chimps. Just think of us rushing frantically from dwelling to dwelling or Starbucks to Starbucks trying to comb the hair of our hundred and fifty closest friends. We would surely all have to have buzz cuts; without one, a person like this girl I knew once, whom I'll call Tawnee, would require for her astounding thick hair about three delightful hours all by herself. But, Dunbar says, precisely because we can't possibly manually delouse all those people—chimps spend 20 percent of their time combing and inspecting one another—we developed conversation. You can converse with four or five people at the same time, and you can also, if need be, simultaneously sharpen arrowheads, cement the stones of a castle keep, or drive a car, depending on your era.

However it came about neurologically and evolutionarily, communicating with one another for no immediate reason has to be the most quintessentially and exclusively human of all our behaviors. According to Emler, the average person of our time spends about 80 percent of his or her life in the company of other people and between six and twelve hours every day talking to those other people. He has also found that self-disclosure accounts for two-thirds of those six to twelve hours, and he breaks down that large symphony of talk into smaller melodies: about 10 percent for states of mind ("I fear Greeks, even when they bear gifts"—Virgil, The Aeneid) or body ("My head is bloody, but unbow'd"—William Ernest Henley, "Invictus") and allotments for preferences, plans, and, the largest of all in this self-disclosure catalog, what Michael Gazzaniga, the neuroscientist and author of Human, calls "doings" ("I found my thrill / On Blueberry Hill"—Fats Domino, "Blueberry Hill"). That is, narrative. That is, storytelling. Finally, in this conversation quantification derby, back to Professor Dunbar, who tells us, in Grooming, Gossip, and the Evolution of Language, that men and women speak about an equal amount—could you have fooled me, or what?—but that men spend two out of every three of their conversational minutes talking about themselves, whereas women spend only about a third of the time doing so.

That makes for an exemplary pie chart, maybe, but it and many other of the conversational data seem suspect to me. For example, assuming that what passes for arithmetic in my brain is right, if two-thirds of everyone's conversation consists of self-disclosure, how can it be that women talk about themselves only a third of the time? Or something like that. More puzzling still is the certitude with which these researchers appear to have decided what qualifies as talking about oneself vs. talking about others. Take the example Gazzaniga cites of a woman—a woman on the catty side, perhaps—who, he says, is talking about someone else: " 'And the last time I saw her, I swear she had gained twenty-five pounds.' " Is that person really talking about the other person? Doesn't her proclivity for such observations, assuming that she has such a proclivity, say more about herself than it does about the weight gainer? Isn't everything we say—at least beyond "Turn right," "What's up?" and "Next!"—self-disclosing? I think so. As Agatha Christie said, "There is nothing so dangerous for anyone who has something to hide as conversation! A human being… cannot resist the opportunity to reveal himself and express his personality which conversation gives him. Every time he will give himself away."

This particular kind of question is particularly important! Because it indicates the complicated and enmeshing nature of good conversation—the lovely pot-holder-esque undistentangleability of it. And those qualities are what often make conversations so gratifying, I think, and what also make them a kind of art. Some are simple and basic, like a child's finger painting. Some have more texture, like a proficient still life. And many are rich and complex, like a Vermeer. Like art, they all have in common the absence of direct utility. Some parts of a conversation usually do have some immediate application, and, as I hope to convince you, conversations definitely have noble benefits that transcend the merely pragmatic. But their most important apparent ingredient—one produced by curiosity, reverie, humor, and playfully associational thinking—is aimlessness, in that word's most neutral definition. Maybe I should write it this way: aim-less-ness. In his engrossing book Conversation: A History of a Declining Art, Stephen Miller gives conversation a similar (if in my view slightly too negative) definition: "Talk without purpose." It may be aimless, but it's not purposeless, as Professor Dunbar has tried to demonstrate evolutionarily and as I hope to show presently.

In an acquisitions meeting at a publishing company whose identity I'll disguise by calling it Camouflage Books, my colleagues and I were discussing how much to offer for a book proposal that claimed to shed new light on the assassination of John F. let's say Shmennedy. One of the people at the meeting (guess who) said about Lee Harvey Shmoswald's hiding place, "You know, I've always been curious about what a 'book depository' actually is." Another person—an excellent firecracker of a publisher with a no-nonsense demeanor—said, "Well, why don't you just hold a seminar on that subject after working hours, Shman? Right now we're trying to figure out what to bid on this project." Shman must have thought, Ouch! No aim-less-ness in stock there. And therefore no conversation, either.

That episode dramatizes the seemingly inevitable conflict between conversation and almost all forms of "results"-oriented human discourse. It's this antagonism—especially in our hyped-up, sound-bite-ridden, profit-hungry global culture—that Stephen Miller and many sociologists hold responsible for the reduction in time we spend conversing. (And reading.) Especially, as Miller laments, in the United States, which, with its pioneering and commercial-traveling and river-damming and belching-smokestack and generally intensely pragmatic and often John Wayne–like ethos and history, has always cast a more critical eye on aim-less social intercourse than have many other modern societies.

Speaking of history—as I will try to do, briefly and with some futility, in the next chapter of this book—since the 1960s, starting with the work of Harvey Sacks, we have in modern times formalized the study of conversation in a sociological discipline called "conversation analysis." That the practitioners and observers of this discipline have abbreviated it to "CA" speaks V's to me about its generally dry tone. Here are some books and articles you can consult to see if I'm right:

Structures of Social Action: Studies in Conversation Analysis

Garfinkel and Ethnomethodology

"Phonetic Detail and the Organisation of Talk-in-Interaction"

Doing Phonology

"Agreeing and Disagreeing with Assessment: Some Features of Preferred/Dispreferred Turn Shapes"

Sequence Organization in Interaction: A Primer in Conversation Analyis, Volume 1

Pretty daunting—though I like the sound of Doing Phonology. I've always had a weakness for the gerundive titles of yesteryear—Raising Arizona, Letting Go, Saving Private Ryan, Driving Miss Daisy, "(Sittin' on) The Dock of the Bay." In many examples of CA you will find discussions not only of turn-constructional units (TCUs; I'm not kidding), but of turn allocational components, sequence organization, adjacency pairs, and preference organization.

Other terms refer not only to mumbling and misspeaking, but to gaffes—who fixes them and how. Gaffes can destroy a conversation more violently than anything else. I have a very good friend who has a large red Gorbachev-class birthmark on her cheek—it's very noticeable. She visited me in the country once, and the first morning, while she was sitting at the dining room table and I was holding a frying pan and serving her bacon that was still sizzling, I said to her, "Lean back a little or you might be permanently disfigured." The discipline of conversational analysis often uses the term "repair organization" for times when conversations break down and are restarted. The repair organization must have been on a break that morning of the bacon.

We can chuckle at the reconditeness of its terms, but CA has sometimes proven helpful to understanding and even improving some important kinds of conversations—like those between doctors and patients, journalists and their subjects, attorneys and witnesses in court, and family members. It has even been applied with some interesting results to the Nixon White House tapes, or so I've read. The tapes usually involve language so opaque that you might think the participants were deliberately disguising their meanings in some weird code:

H. R. "BOB" HALDEMAN: We had [Charles "Chuck"] Colson make a…

PRESIDENT NIXON: Chuck is something else.

HALDEMAN: Yeah. You know, [Edmund] Muskie sent those oranges down to the veterans—that group on Saturday, I mean.


HALDEMAN: He didn't go down himself, but he sent oranges.

PRESIDENT NIXON: Did Colson order some oranges for him?

HALDEMAN: Colson sent oranges out yesterday. [laughs] From Muskie.


HALDEMAN: I don't know whether it's out yet or not. They'll get it out. [laughs]

PRESIDENT NIXON: He just ordered them?

HALDEMAN: [laughs] Yeah. An awful lot of cases of oranges. I don't know how the hell he does that stuff, but he… It's good, you know, he's been around the District here so long, he has a lot of contacts, and he, as a local guy, he can get stuff done here that… And he's got no—He's going to get caught in some of those things [unclear].

PRESIDENT NIXON: [unclear] Well, he has been caught.

HALDEMAN: [unclear] And he has been caught.

PRESIDENT NIXON: It's all right.

Medical or political, even in raw transcripts of basically banal exchanges there have to be many layers of unspoken subtext. My point is that like works of art, and most human transactions, anything more than truly rudimentary talk ("Have some?" "Yes, please") has innumerable nuances, including what the observer himself brings to it—a sketchy knowledge of the language being spoken, for example, or being heartbroken, or in love, or sleepy, or drunk. Even video recordings of a conversation, which can of course show some body language, can't reproduce the event's entirety. None of the subjects' pheromones waft through the air, the people look five to ten pounds heavier than they are, the colors aren't true. And, as is the case with our real-life experience, no amount of transcription, recording, or even being present—or even MRIs—can tell us with any precision what is going on in another person's mind. Or in our own, really, for that matter. And these latencies are surely important ingredients of any conversation.

So maybe the best way to understand what's going on when we talk to one another, and to get better at it and enjoy it more, is to approach this exclusively human activity the way a good critic approaches any other human artifact that has inexhaustibly various aspects. That's part of what I'll do here, in the hope that it will interest, entertain, and benefit readers—partners in the conversation that any writer is implicitly having with them, even if those readers can answer only silently. After a brief look at the history of conversation, I'll address its components through an examination of a single specific recorded conversation, held over a meal between me and someone I know only slightly. I paid for the meal. Actually, I paid for five meals, four of them not with me as a participant, but one of the first recorded conversers pulled out of the deal later on because he thought he had divulged too much—especially about one of his wife's obstetricians. I'll discuss better and worse ways to think about and approach these conversational components and try to show that there is a "deep structure" that underlies many conversations. The main focus here will be on those situations in which we talk for an hour or so to someone we don't know very well or at all—at a party, on an airplane (two very clearly consenting participants required!), at a bar, and on many other kinds of occasions, both planned and unplanned. My partner in this conversation and I are both writers, but there were many differences between us as well—age, gender, background, religion, and so on. And I believe that the way our talk went is, however loosely, the way most spontaneous talks go, barring extreme disparities in the personal circumstances—wealth, poverty, mental illness, and so on—of the talkers. (By the way, I think airplane passengers should all be asked to wear something like Indian caste marks on their foreheads—red for "No talking, please," green for "Yes, you can tell me about your cousin's goiter, and I will tell you about Tawnee, who broke my heart." Or little face icons, maybe, with zipped or unzipped lips. I've been passionate about the issue of airplane conversations ever since I more or less conquered my fear of flying and too soon afterward sat next to a woman who for more than two hours told me why she was so afraid of flying.)

The underlying conversational structure I've mentioned consists—shockingly, I realize—of a beginning, middle, and end. But, as I hope chapters 3 and 4 will show (two chapters only because one would be too long), it's more topic-specific and interesting than that. Greetings are sometimes fraught: After all, the handshake grew out of the custom of demonstrating that one isn't carrying a concealed weapon. Fraught example: A fairly good young friend recently came to my house for dinner and noticed that I was wearing a (very manly!) bracelet—a recent addition to my possessions, much less expensive than your usual late-middle-age, jaunty-sports-car-convertible, death-denying protest vehicle. "What's that, Dan?" he said as we showed each other that we didn't have concealed weapons. "It's a bracelet," I said. "I thought it was a watch at first," he said with disappointment. "No, it's a bracelet," I said. "Well," he said, without a trace of humor or irony or any other hint of mitigation, "it makes you look like a fag." So it might be interesting and helpful to take a closer look at the way hellos work.

The main part of a conversation, particularly with a new acquaintance, establishes some kind of common ground. In the rare event that it doesn't, the whole meeting will be something of a failure. Finding that common ground and then moving beyond often has two stages. The first is the Survey, in which the people involved discuss some important aspects of their identities. The second stage is Discovery, in which sometimes significant connections are unearthed: "We have the same middle names!" "My kids don't get along very well, either." These, again, can be trivial or revealing: "I have flat feet," "I had a one-night stand two days before I got married—two days minus one day, to be completely honest." This second phase often increases the sense of trust between the two conversants, and that trust in turn leads them to take Risks.


On Sale
Jan 4, 2010
Page Count
240 pages

Daniel Menaker

About the Author

Daniel Menaker has been a part of America’s life of letters for almost forty years. As a writer, he has met and talked to thousands of people about their work and their lives. He is widely read and well versed in psychological literature and practices and, as an editor at Random House, has had countless meetings and other exchanges with writers, agents, public figures, and ordinary people. His own writing has appeared in the New Yorker, the New York Times, and Slate; he lives in New York with his wife and their two children.

Learn more about this author