Stephen Sondheim: Examining His Lyrics And Life.

February 16, 2012

Guests: Josh Putnam-Stephen Sondheim

TERRY GROSS, HOST: This is FRESH AIR. I'm Terry Gross. While following the Republican primary race this year, I found myself wondering why the voting system varies from state to state. Why do some states have caucuses and others primaries? And if a caucuses' results aren't binding, how are the delegates actually chosen?

To get some explanations, we invited Josh Putnam, who runs the elections blog Frontloading HQ. He's a visiting assistant professor of political science at Davidson College in North Carolina. His focus in his research and in the classroom is on campaigns and elections.

Josh Putnam, welcome to FRESH AIR. Why are there so many different systems for selecting delegates, state by state? Does the Constitution or federal law have anything to say about that?

JOSH PUTNAM: Well, the Constitution doesn't really say much of anything. If we go back to the discussions around the founding, you know, it doesn't even - the Constitution, that is - doesn't even account for parties in there. So this is a situation where the states have the authority, one of those enumerated powers that's left up to the states to decide.

GROSS: So who does decide what happens in the state primary process? Is it the state? Is it the party?

PUTNAM: It's a combination of the two. Ultimately, the state party has the final say in whether or not they're going to have a primary or caucus. But at the end of the day, there's also this incentive, I suppose, that state parties have to opt into a presidential primary, simply because the state funds it in most cases.

Otherwise, they're footing the bill for a caucus or a primary of their own, which can end up being fairly expensive.

GROSS: So if it's a primary, then the state pays, and if it's a caucus, then the party pays?

PUTNAM: That's right.

GROSS: So what's the difference between a primary and a caucus?

PUTNAM: Well, a primary functions pretty much like a general election does. You go in, you sign in with an election official, and you go and cast your vote and leave. A caucus is different from that. You go in - and certainly the presidential preference vote is part of the process - but we're talking about, in addition to that, a three- or four-hour meeting that deals with other party business: setting up the platform process for the subsequent state convention and so on and so forth. So again with a caucus, we're dealing with something that's much more of a time commitment on the part of caucus-goers in this case.

GROSS: And so that's one reason why there might be a much smaller turnout for a caucus than a primary?

PUTNAM: That's a significant reason why we see lower turnout.

GROSS: So let's look at the caucuses in Maine that were held earlier this month. So Mitt Romney won over Ron Paul, 39 percent to 36 percent, and it was a less-than-200-vote difference, and at the same time, there were votes in three counties that weren't counted. So if those votes were counted, who knows what the results will be.

But putting the confusion of those results aside for a moment, let's just try to understand how those Maine caucuses work and what the votes actually mean. What do those votes count for? Was that just a straw poll? What do those votes actually mean?

PUTNAM: The problem this time around, and perhaps with the Republican process, is that these are non-binding votes. I mean, we've heard this over and over again in the news, from Iowa on through the last few caucus contests that we had - Maine over this past weekend - for instance, that again, this is a non-binding contest.

And that's true. Typically, folks will come in, they'll have this president preference vote, a straw-poll vote, and then people have the option of leaving after that. But attendant to that is a delegate selection process, where folks are selected or elected by those that are still there, to move on to the next round, whether it be a county caucus or a district caucus or even on to the state convention.

What we're talking about, a large group of people at the precinct level choosing a smaller group of people to move on to the county level, and then the process repeats itself. And the delegates gradually winnow themselves down. So they're choosing delegates from among the people that are left at the county level, to move on to the district level, at the district level to move on to the state convention.

GROSS: So first there's a straw-poll vote, that doesn't count, but then people stick around and actually vote for the delegates?

PUTNAM: Right, right. So it's naïve of us to think that there's no transference of presidential preference through this process. But we don't have, at least with the Republican presidential process in 2012, a firm grasp on how that's actually translating from one step to the next.

GROSS: Let's apply this to what the Ron Paul strategy seems to be now. The Ron Paul strategy seems to be that after the vote in a caucus, the Ron Paul people stick around so that they can have more sway in the actual selection of delegates process. Explain how that works. Like, explain what they're trying to do, what their strategy is and how it would work if it succeeded.

PUTNAM: Well, again, the assumption going into this is that the number of delegates moving forward, are going to be close to proportional to vote in the straw-poll vote. But there's nothing in most of these state party rules that requires that. And that's the kind of loophole in the process that the Paul folks are taking advantage of, that they will stick around and be very regimented in making sure that their supporters gobble up all those delegate slots to the next round of this process.

And that very greatly increases their chances of pushing folks through to the national convention. The straw poll is only for show, essentially. And really where this comes in is that in most cycles, the nominee is settled on really quickly. So the straw poll becomes a part of a candidate's momentum, for instance, and then once we finally get around to the state convention, we already know who the nominee is and those delegates line up behind that nominee.

At this point in 2012, we've got a situation where we don't have a clear front-runner, we don't have a clear front-runner with momentum, so we've got a competition for these delegates, and the Paul folks are trying to exploit that.

GROSS: So Rick Santorum won the caucuses in Colorado, in Minnesota, you know, recently. Does that mean he won the delegates? It doesn't then, huh?

PUTNAM: It does not necessarily mean that he won the delegates, no. To prove that, we'd need to have a handle on what his organization was like across the state, and we just don't have a feel for that. We don't have a feel for that for what's happening with Ron Paul, either.

GROSS: There's been several states that have had pretty serious problems in the Republican primary and caucus process this year. There was Iowa, Romney was announced as the winner, a few weeks later they announced that Santorum was probably the winner, but there were problems in several districts, there's never going to be a completely accurate count.

And the in Missouri, why don't you explain what happened in Missouri.

PUTNAM: In terms of having both a primary and a caucus?

GROSS: Yes.

(SOUNDBITE OF LAUGHTER)

PUTNAM: Oh, that may take several years to explain that. Essentially, the Missouri state legislature had a very difficult time attempting to move back its presidential primary to comply with the new Democratic and Republican party rules governing when states could hold primaries and caucuses.

And they couldn't do it. There were several bills that made their way through the legislature, and one was passed that would have moved the process back to March 6th, Super Tuesday, but it also contained a provision that would have stripped the governor of the power to appoint folks to vacancies to statewide office.

Because of that, the governor vetoed the bill, which scrapped the move to March 6th, and in a subsequent special session, the legislature was unable to get on the same page as far as moving it back. There was a split within the Republican Party that's in control of the legislature there.

Part of the legislative caucus there wanted to keep the primary early so they could have influence despite the penalties that would have been attendant to that, and the other part was more pragmatic, I suppose, about it, wanting to move back to maintain the full allotment of delegates that they had.

As I said, couldn't get on the same page and could not even cancel what would have been in there, and what was, by all accounts, a meaningless primary that took place on February 7th.

GROSS: Because the party is not going to count what happened on February 7th.

PUTNAM: It was a straw poll vote that's even less meaningful than what we're seeing in some of these caucus states, yeah.

GROSS: So Santorum won this straw poll on February 7. What does that mean? I mean, he seemed so victorious, like he won it, but what does that actually mean?

PUTNAM: Well, it, in conjunction with the wins in Minnesota and Colorado that night, showed that he had some momentum in this race and that he was a viable alternative. Why Missouri was so important was because Newt Gingrich didn't make the ballot. So it was a clear demonstration, though it was completely meaningless, that Rick Santorum could hold his own against Mitt Romney and really dominate the contest.

GROSS: So how has the primary process been undercut by the errors in calculations, the votes that weren't tabulated? It just seems to be a lot of problems this year.

PUTNAM: There seemed to be some problems this year. I think a part of that has to do with the calendar that developed this time around is that we had, you know, just this very slow buildup of contest, and we haven't had, you know, that Super Tuesday yet, where we've got 10, 12 contests happening on the same day.

And being that as it is, we've been able to really scrutinize the votes that are taking place in each state, one by one.

GROSS: Having this overview of how, state by state, the delegate selection process varies, it just makes you wonder, does everybody's vote count equally in the primary process? And in some ways, it seems like the answer is no because people are voting in so many different kind of systems. Like one system's a straw poll, and one system is a primary, where your vote goes directly to choosing the delegates. So do you think of the votes as all being equal?

PUTNAM: Well, I guess the main problem there, at least traditionally, has been for states that hold either primaries or caucuses after the point at which someone has passed that 50-percent-plus-one delegate barrier, that they don't matter. There's no incentive for voters in those states to come out and vote for a candidate of their preference, other than to affect the delegate selection process that will obviously put out delegates that will support who's going to be the nominee at that point.

But yeah, it's fair to say, potentially, that given the rules that it's - votes count differently in different states, or at least voters have different incentives, that it's easier to participate in a primary state than it is in a caucus state. The time commitment alone forces folks away from that process.

GROSS: But I guess there was a time when people didn't vote at all, when it was really just backroom deals, choosing the candidates.

PUTNAM: That's it, and, you know, those pre-'68 days, or I guess '68-and-before days, you know, the primaries and caucuses were similar to what we saw in Missouri recently: They were just for show, to show that the candidate had some support in that state and that, you know, potentially they'd do well there in a general election and to make the case for folks that were actually making these decisions at the conventions later on.

GROSS: Kind of like a test screening, get audience reaction.

(SOUNDBITE OF LAUGHTER)

PUTNAM: That's right, that's right.

GROSS: So let's go back to 1968, when the primary process was reformed, after the 1968 Democratic convention that nominated Hubert Humphrey. And just to set the scene a little bit, this is the convention in which there's a huge anti-war protest outside of the convention. The protestors are being tear-gassed and beaten. Journalists are being gassed and in some cases beaten by police.

And inside, a very chaotic scene, as well. So let's go inside the convention hall. Who were the candidates?

PUTNAM: Who were the candidates? So again, we'd had a process where Lyndon Johnson had lost the New Hampshire primary and opted to not seek re-election in '68. That led to Eugene McCarthy being a candidate. He'd been on the New Hampshire ballot, had come in very close to Lyndon Johnson there.

Robert Kennedy had been a part of the process, as well, obviously was assassinated just after the California primary in June of that year. So wasn't obviously a viable candidate at the convention, but also the person who emerged late in the process, who had not participated in any of these presidential primaries, was Hubert Humphrey, the eventual nominee.

And that was viewed in light of the chaos going on outside of the convention, was viewed as not being necessarily and accurate reflection of what rank-and-file members of the Democratic Party in 1968 thought their nominee should be or who they thought it should be.

GROSS: How did Humphrey get to bet the nominee if he hadn't participated in any of the primaries?

PUTNAM: Well again, the primaries at that point in time were just a show. They were a way for candidates to demonstrate that they had some strengths and viability in any given state or region that could translate to the general election. The ultimate decision was not bound in any way by - I'm sorry, the ultimate decision for the nominee, for the nomination, was not bound by anything that had happened in those primaries and caucuses at that time. So they merely served as advisory.

And in '68, that advice was essentially ignored because Humphrey had not participated in any of the primaries.

GROSS: And what were some of the reforms that came out of this?

PUTNAM: Well, the major thing was that it made, unlike what we'd seen in the time in '68 and before, it made the primaries and caucuses binding, that the decisions made in those primaries and caucuses would affect the delegates that were chosen to go to the convention and thereby affecting who the ultimate nominee was going to be.

GROSS: My guest is Josh Putnam. He runs the elections blog Frontloading HQ and is a visiting assistant professor of political science at Davidson College in North Carolina. This is FRESH AIR.

(SOUNDBITE OF MUSIC)

GROSS: We're talking about the primary process with Josh Putnam. He runs the elections blog Frontloading HQ. What has surprised or interested you most in this current primary process?

PUTNAM: Surprised, I guess I'm most surprised at the volatility that we've seen on the Republican side this time around, that it's not something that we've typically seen. I mean, this is just a very, very atypical cycle for the Republican Party, that typically Republican primary voters will line up behind a nominee fairly quickly, and we're just not seeing this, this time around.

And that's just indicative of the divisions within the party right now, that you've got on the one hand a Tea Party faction, on the other an establishment faction within the party that's trying to decide who the best candidate is to go up against Barack Obama in the fall.

GROSS: When you look at, like, the number of debates that the Republicans have had so far, a number I frankly have lost track of, and think of all the money that the candidates spend on this really long primary process, you know, it kind of makes you wonder why not make the primaries and the caucuses all on one day or a couple of days or have them in, like, one or two months. Why is dragged out for so long? What's the logic behind that?

(SOUNDBITE OF LAUGHTER)

PUTNAM: That may be a question better asked of the national parties. I mean, again, what we've got here is something of an ad-hoc system that's developed over more than a generation, right. But the sequential nature of it has its benefits.

It on the one hand can allow the vetting of the eventual nominee, that all the warts can be dealt with before the general election. But alternatively, it can allow for lesser-known candidates to rise up and take advantage of the process and maybe emerge as a nominee, both of which we would, kind of, lose if we just had a one-shot-deal national primary.

GROSS: You not only study the primaries, I mean, you study the whole election process, including redistricting, the Electoral College. So when you stand back, and you look at the primaries, and you look at the actual election process, do you ever think our democracy is maybe slightly less democratic than we think it is in terms of all votes being equal?

PUTNAM: That's a tough one. I just really try to steer clear of normative questions like that, these would've, could've, should've sorts of questions. Because when we talk specifically about the presidential primary process, at the end of the day, it's party business, and they can decide how democratic it is or how seemingly authoritarian it seems to be or whatever, right, the party has a heavy hand in it.

But this type of scrutiny that we're seeing in 2012 is a healthy exercise, right, that we're looking at, very closely, how this caucus process is run. And we'll leave it up to the interaction between the rank-and-file members of any given party and, you know, the state party, the national party itself to determine whether or not there need to be some sort of tweak to the system to make it more reflective of our or their collective vision of what democracy is.

GROSS: Well, Josh Putnam, thank you so much for talking with us.

PUTNAM: Glad to be with you, Terry.

GROSS: Josh Putnam runs the elections blog Frontloading HQ, and he's an assistant visiting professor of political science at Davidson College in North Carolina. I'm Terry Gross, and this is FRESH AIR.

TERRY GROSS, HOST: This is FRESH AIR. I'm Terry Gross. Stephen Sondheim's musical, "Merrily We Roll Along," ran for only 16 performances after opening on Broadway in 1981. So, for fans of the show like me, a revival is cause for celebration. The show is currently being performed for a two-week run as part of the Encore! series "Great American Musicals in Concert" at New York City Center.

We're going to listen back to excerpts of two interviews from 2010, in which Sondheim talked about that show and his songwriting process. "Merrily We Roll Along" is about three good friends, two young songwriters and a fiction writer, who are eager to make their mark. The story shows how they change over the course of 25 years. One character, composer Franklin Shepard, heads in a commercial direction, which ruins the friendships and turns him cynical and self-centered.

The story is told chronologically in reverse, so it starts with bitterness and ends with the idealism of youth. The best-known songs from the show are "Old Friends," "Not A Day Goes By," and "Good Thing Going."

Stephen Sondheim, welcome back to FRESH AIR. You say that people assume that a lot of your songs are really autobiographical, but they're not, with the exception of "Opening Doors," from your 1981 show "Merrily We Roll Along." And I saw a revival of this show a few years ago. It didn't last long, sadly, on Broadway, but I saw a revival by the York Theater Company.

STEPHEN SONDHEIM: Yeah.

GROSS: It was wonderful. I loved the show, and I loved the songs from the show. So I want to play "Opening Doors," and then I want to talk a little bit about it.

SONDHEIM: OK.

GROSS: And do you want to describe where the song fits into the story?

SONDHEIM: The idea of the show was to go - as the listeners may not know - go backwards in time from a very successful group of 40-year-olds or 45-year-olds and take them back to their very youthful days before they compromise their principles. It goes backwards in time. It's based on a George Kaufman, Moss, Hart play.

The song takes place over a period of two years in the lives of the three leading players. Two of them are songwriters, a lyricist and a composer, and their best friend is a woman who is, a young woman who is a budding novelist. And it's the three of them trying to break into, well, the two guys into show business, and she's trying to finish writing a book.

GROSS: So this song is at the point where they're kind of hoping to become real, you know, a real...

SONDHEIM: That's right. Yeah. They're...

GROSS: ...composer, a real lyricist and a real novelist.

SONDHEIM: They're opening doors. They're knocking on doors. They're knocking on doors.

GROSS: They're opening doors. And we're going to hear this sung by, in the original cast recording, sung by Jim Walton, Lonny Price and the part of the producer, who interjects in the middle here, will be sung by Jason Alexander, who played George on "Seinfeld."

SONDHEIM: Right.

GROSS: So here we go, from Stephen Sondheim's "Merrily We Roll Along."

(SOUNDBITE OF SONG, "OPENING DOORS")

(SOUNDBITE OF TYPING)

JIM WALTON: (As Franklin Shepard) (Singing) Bum-bum da-da-da-da-da. Bum-bum-bum, bum-bum, bum-bum.

(As Franklin Shepard) How's it coming?

LONNY PRICE: (As Charley) Good. You?

WALTON: (As Franklin) (Singing) Good.

PRICE: (As Charley) (Singing) One minute.

(SOUNDBITE OF TELEPHONE)

WALTON: (As Franklin) (Unintelligible), Mary...

PRICE: (As Charley) Say hello.

ANN MORRISON: (As Mary) I think I got a job.

WALTON: (As Franklin) Where? What's that?

MORRISON: (As Mary) Thank you. Writing captions.

WALTON: (As Franklin) What about the book? Did you give the publisher the book? Good.

MORRISON: (As Mary) No.

WALTON: (As Franklin) Mary.

PRICE: (As Charley) I finished.

WALTON: (As Franklin) Let me call you back.

PRICE: (As Charley) This is just a draft.

WALTON: (As Franklin) Right.

PRICE: (As Charley) Probably it stinks.

WALTON: (As Franklin) Right.

PRICE: (As Charley) I haven't had the time to do a polish.

WALTON: (As Franklin) Will you sing?

PRICE: (As Franklin) Right. (Singing) Who wants to live in New York? Who wants the worry, the noise, the dirt, the heat? Who wants the garbage cans clanging in the street? Suddenly I do.

(As Franklin) They're always popping their cork - I hate that line - the cops, the cabbies, the salesgirls up at Saks, you gotta have a real taste for maniacs. Suddenly I do.

JASON ALEXANDER: (As Joe) (Singing) That's great. That's swell. The other stuff as well. It isn't every day I hear a score this strong, but fellas, if I may, there's only one thing wrong:

(As Joe) (Singing) There's not a tune you can hum. There's not a tune you go bum-bum-bum-di-dum. You need a tune to go bum-bum-bum-di-dum. Give me a melody.

(As Joe) (Singing) Why can't you throw 'em a crumb? What's wrong with letting 'em tap their toes a bit? I'll let you know when Stravinsky has a hit. Give me some melody.

(As Joe) (Singing) Oh sure, I know, it's not that kind of show, but can't you have a score That's sort of in-between? Look, play a little more, I'll show you what I mean:

PRICE: (As Charley) (Singing) Who wants to live in New York? I always hated the dirt, the heat, the noise. But ever since I met you, I...

ALEXANDER: (As Joe) (Singing) Listen, boys, maybe it's me, but that's just not a hum-umam-umam-umamable melody. Write more, work hard, leave your name with the girl. Less avant-garde, leave your name with the girl. Just write a plain old melodee-dee-dee-dee-dee-dee - dee-dee-dee-dee-dee-dee...

GROSS: That's "Opening Doors" from "Merrily We Roll Along" by my guest, Stephen Sondheim, who said this is his really autobiographical song. So is the part autobiographical where the producer complains that it's not a song you can hum, give me a melody?

SONDHEIM: Oh, sure.

(SOUNDBITE OF LAUGHTER)

SONDHEIM: Oh sure, oh sure, oh sure, sure. But it's autobiographical in the general sense. It's, you know, first of all, I didn't have a collaborator. I mean, it's not specifically autobiographical.

I wrote my own lyrics and my own music, and the girl is merely an amalgam of people like, particularly I was very close to Mary Rogers, Dick Rogers' daughter. She became a composer, as well as a novelist, as a matter of fact, and of course, Hal Prince, who was a producer and then eventually a director.

And we were all very close to each other. It's not specifically based on us, but it's on the ambience of our lives and the speed and the excitement and the disappointment and the triumph, et cetera.

The whole business of hummability, of course, has to do with familiarity. If you hear a tune enough times, you'll hum it. You know, you can - first time I heard the Berg violin concerto I thought, what is this noise? And the third time I heard it, I thought, oh, that's interesting. And the fifth time I heard it, I was humming along with it.

And I remember being at the intermission of "A Little Night Music" when it first came out and hearing somebody say oh, that "Weekend in the Country," that's such a catchy tune. Well, you know, very few people accuse me of writing catchy tunes, and of course it was a catchy tune. She just heard 11 choruses of it, and so of course she could hum it.

GROSS: Now, the producer sings: I'll let you know when Stravinsky has a hit, he's saying sarcastically.

SONDHEIM: Mm-hmm.

GROSS: Now, you studied with the avant-garde composer Milton Babbitt.

SONDHEIM: Mm-hmm.

GROSS: When you studied with him, was your ambition Broadway, or was it more...

SONDHEIM: Oh, no, I always wanted to write songs. Well, he's a songwriter manque. I wanted to learn compositional technique, and that's what I learned from him.

But we would spend - we had four-hour sessions once a week, and we would spend the first hour analyzing songs by, oh, Jerome Kern or by de Sylva, Brown, and Henderson, the classic songs of the American theater and American movies.

And we spent an hour, you know, on, you know, songs, and then three hours on Beethoven and Bach. And it was all about essentially compositional analysis. But no, I only wanted to write songs. I didn't want to write concert music.

GROSS: Can you give me an example of an insight you got from Babbitt studying, say, a Jerome Kern song?

SONDHEIM: One of the things we analyzed in detail, one of the songs, was "All the Things You Are," which has a remarkable harmonic structure in it, which among other things consists of the fact that the tonic chord isn't played until the end of the song, and it goes from a circle of fifths and then breaks the circle of fifths with a tritone, which echoes itself not only in the melody but also in the bass and defines both the key that the song is written in and the key to which it's going, et cetera, et cetera, et cetera.

I've actually reproduced that hour-long analysis he gave me to students I had at Oxford when I taught at Oxford. And it's lodged in my mind because it is a way of approaching, when you are trying to hold a song together, how you hold it together harmonically and still make it fresh. Kern was a master at that.

GROSS: My guest is Stephen Sondheim. We'll hear more after a break. This is FRESH AIR.

(SOUNDBITE OF MUSIC)

GROSS: The 1981 Stephen Sondheim musical "Merrily We Roll Along" is being revived for a two-week run by the Encore! series "Great American Musicals in Concert" at New York City Center.

We're listening back to excerpts of two interviews with Sondheim, in which he talked about that show and his songwriting process. Let's hear another song from the show. This is "Good Thing Going." Now this is the finished version of the song we heard excerpted earlier within the song "Opening Doors," when two characters perform their song for a Broadway producer who rejected it as not hummable. We'll hear Adam Heller and Malcolm Getz, who starred in the 1994 revival of "Merrily" by the York Theatre Company.

(SOUNDBITE OF SONG, "GOOD THING GOING")

ADAM HELLER: (As Charley Kringas) (Singing) It started out like a song. We started quiet and slow, with no surprise. And then one morning I woke to realize we had a good thing going. It's not that nothing went wrong. Some angry moments, of course, but just a few, and only moments, no more. Because we knew we had this good thing going.

(As Charley Kringas) (Singing) And if I wanted too much. Was that such a mistake at the time? You never wanted enough - all right, tough, I don't make that a crime.

GROSS: Let's get to this idea of opening doors. What were some of the first doors you knocked on before actually getting to Broadway and writing lyrics for "West Side Story"?

SONDHEIM: Well, I played for an awful lot of people. I remember once playing for a guy named Cy Feuer, who was one of the producers of "Guys and Dolls." He had also been a musician and was head of the music department at Universal. And I remember he criticized me for having too many B-flats in a melody. I remember he said that. And I thought, gee whiz, what is he talking about?

(SOUNDBITE OF LAUGHTER)

SONDHEIM: I mean, you know, he wanted to show me that he knew a lot about music, is what it was. And he might have been right, but I don't think he was. And I played for a number of producers and directors and generally was dismissed. It was, you know, I snuck in through "West Side Story" where, you know, they were the big guys there, Leonard Bernstein and Arthur Laurents and Jerry Robbins. So it was their problem to get the show and believe me, it was not an easy show to get on.

GROSS: Did you learn anything working with Bernstein and watching him work?

SONDHEIM: Oh, sure. A great deal. Yes. Mainly I learned something about courage. I learned – Lenny was never afraid to make big mistakes. He was never afraid to fall off the top rung of the ladder and I learned by implication that the worst thing you can do is fall off a low rung. If you're going to make a mistake, make a huge one.

GROSS: But you've talked about George Gershwin and Harold Arlen as great influences on you and they were both very influenced by jazz. Did you listen to much jazz or pop when you were in your formative years?

SONDHEIM: Nope. Nope, I didn't and I'm not very influenced by jazz. First of all, the whole idea of jazz is improvisation and instrumentalists and because I'm only a piano player and have never played in a band (technical difficulty) feeling for that. Also, I think by nature I'm too conservative. I'm just – I only improvise at the piano when I'm writing a song but I never improvise for anybody else or in front of anybody else or at a party or anything like that.

And I don't think I would be good at it. I'm much too constrained. It's partly my training. My first music teacher who was a professor at Williams College was a very, very kind of Mary Poppins kind of teacher with, you know, he laid down the rules. And that appealed to me a lot, the idea of rules of how you write music that say what music consists of.

That it's not just sitting and waiting for an inspiration but that you take a melodic idea that you have that might be an inspiration but then you develop it and you work with it and work it out. You don't just fiddle around at the piano to do it. And that appealed to me a lot but that's very conscious composition and that's also what I studied with Milton Babbitt.

And that is reverse of jazz. In fact, it's always struck me so odd that Mitlon, who is so knowledgeable about composition and composes according to a set of rules, some of which he makes up himself, is also such a jazz fan. I can never put those two things together. At any rate, no, I was not influenced by jazz. I was influenced by Gershwins' and Kearns and Arlen's songs and particularly by their use of harmony.

GROSS: There's a beautiful song in "Merrily" that's sung twice and I'm thinking of "Not A Day Goes By" and both versions – each version has a different meaning because one's at the beginning of a love affair and the other is during a divorce.

SONDHEIM: Right.

GROSS: Can you talk about writing that song with two different meanings in mind?

SONDHEIM: Well, I wrote the whole score knowing that it was going to go backwards in time and I thought what does that imply? Well, it implies that something that you and I sing today 20 years from now will have a different meaning to both of us. It doesn't have to be that we get divorced. Maybe it'll be memories of something, but everything that happens at a given time in your life has echoes and resonances afterwards. What I would call, like, reprises, really, of thoughts, of moments in your life that happen in different contexts.

So I thought if I'm going to write the show that goes backwards in time, we'll start with the reprises. That is to say, start with the variation on the theme and then go back to the theme. And that's what happens here. It happens with a lot of other songs in the show, too.

But this one very specifically with the lyric because it applies to two very distinct and distinctly defined situations – one a divorce and one when they got married. So you're taking two high spots of their lives, their marriage and their divorce.

GROSS: So we'll hear both versions of "Not A Day Goes By" from the 1994 York Theater revival.

SONDHEIM: That's the way to illustrate it.

GROSS: Yeah.

(SOUNDBITE OF SONG, "NOT A DAY GOES BY")

ANNE BOBBY: (As Beth) (Singing) Not a day goes by, not a single day but you're somewhere a part of my life and it looks like you'll stay. As the days go by I keep thinking when does it end. Where's the day I'll have started forgetting? But I just go on thinking and sweating and cursing and crying and turning and reaching and waking and dying and no, not a day goes by, not a blessed day, but you're still somehow part of my life.

(As Beth) (Singing) And you won't go away. So there's hell to pay. And until I die, I'll die day after day after day after day after day after day after day till the days go by. Till the days go by. Till the days go by.

AMY RYDER: (As Mary) (Singing) Not a day goes by, not a single day, but you're somewhere a part of my life and it looks like you'll stay.

MALCOLM GETS: (As Frank) (Singing) As the days go by. I keep thinking when does it end.

RYDER: (as Mary) (Singing) That it can't get much better much longer but it only gets better and stronger and deeper and nearer...

MALCOLM GETS AND AMY RYDER: (As Mary and Frank) (Singing) ...and simpler and freer and richer and clearer and now not a day goes by.

BOBBY: (As Beth) (Singing) Not a blessed day but you somewhere come into my life and you don't go away.

MALCOM GETS AND AMY RYDER: (Singing) And I have to say if you do I'll die. I want day after day after day after day after day after day after day after day after day till the days go by. Till the days go by.

BOBBY: (Singing) Till the days go by.

GROSS: That's "Not A Day Goes By" from the 1994 York Theater Company revival of Stephen Sondheim's "Merrily We Roll Along." We heard Anne Bobby, Malcolm Gets, and Amy Ryder. Our guest is Stephen Sondheim.

Now, another question about your formative years. You went to a Quaker school but you also went to the New York Military Academy.

SONDHEIM: Yeah, that was earlier.

GROSS: I can't imagine you being a cadet. I know that there was also am emphasis in athletics at the school. I went on the website and it said all cadets must participate in sports throughout the year. So what was it like for you to be...

SONDHEIM: Well, first of all, I went when I was 10 years old.

GROSS: Oh, OK.

(SOUNDBITE OF LAUGHTER)

SONDHEIM: At 10 and 11 years old, only those two years. So, yeah, I think you may have the wrong picture. I mean, it was also because my parents had just divorced and military school was always considered a place to send kids of divorced parents. A lot of my classmates were kids of divorced parents. And it was a lifesaver but your life becomes chaotic suddenly when your parents split up and military school is bringing order to chaos.

You have to be at a certain place at a certain time, you have to polish the buttons on the uniform, you have to parade here, you have to take orders there, and it was wonderful. There's a sense of structure and I think psychologically it must've saved my life.

GROSS: You know, that actually really fits into what you were talking about, about wanting rules and structure in music.

SONDHEIM: Mm-hmm. Yep. Order out of chaos, order out of chaos. That's why I like crossword puzzles. Order out of chaos.

GROSS: Right, right, right.

SONDHEIM: I think that's what art's about anyway. I think that's why people make art.

GROSS: To create order...

SONDHEIM: Out of chaos.

GROSS: ...in a world that's chaotic.

(SOUNDBITE OF LAUGHTER)

SONDHEIM: The world has always been chaotic. Life is unpredictable. There is no form and making forms gives you solidity. I think that's why people paint paintings and take photographs and write music and tell stories that have beginnings, middles, and ends, even when the middle is at the beginning and the beginning is at the end.

GROSS: Stephen Sondheim recorded in 2010. His 1981 musical "Merrily We Roll Along" ends its two-week revival Sunday presented by the series "Encore's Great American Musicals in Concert at New York City Center." This is FRESH AIR.

(SOUNDBITE OF MUSIC)

GROSS: Let's get back to an interview I recorded with Stephen Sondheim in 2010. Can we talk a little bit about your process of songwriting? When you're writing at the piano are you recording what you're playing or are you just like...

SONDHEIM: Never.

GROSS: No taping it?

SONDHEIM: No, no. No, no. Just notating.

GROSS: Mm-hmm.

SONDHEIM: No tape. The process of putting something down on paper is very important, I think, and keeping the stuff alive in your head just – you have to make – even if you're just improvising you have to make little decisions just to put it down on paper.

You can improvise a phrase and as you're putting it down and play it again, you may think, wait a minute, that A-flat, no, doesn't sound right. And you change things as you go along, even though you're just sketching. It's precisely what an artist does when he – a painter does or somebody who draws, when he sketches.

When you look at, you know, the sketchpads of anybody, you know, Michelangelo or Leonardo and see how they experiment with, you know, a horse's head or a hand or something like that. That's precisely the analogy. You're putting, you know, a finger on the – or a hand on the music paper before you try to work out a whole body.

GROSS: Now, when you're working with rhyme what's your process for figuring out options for rhyming words?

SONDHEIM: Oh, well, you use a rhyming dictionary is what you do. And the important thing is to get the thought first, to know what you want to say. And then how you want to phrase what you want to say, and then as the music develops, you'll start to improvise a rhyme scheme or to sense a rhyme scheme.

And then if you're, you know, you say, all right. I've got this line that ends with day and I want to say she loves him. So how will I? And then you go through the rhyming dictionary and rhyming dictionaries are useful for rhymes like day. They're not useful for trick rhymes. Those you just think up, you know.

But there are so many rhymes for day and you want something that will somehow encompass or pinpoint what you want to say - there's a rhyme right there - about this situation. And I use a particular rhyming dictionary called the Clement Wood, which the advantage of which is that all the rhymes are listed vertically instead of horizontally.

So your eye sweeps up and down the page until a word catches it. The problem, for me anyway, with rhyming dictionaries that list things horizontally is that your eye tends, because you start to get impatient, to skip over the words. But when your eye goes up and down a page you don't skip over as much.

And then suddenly a word will pop out and, you know, bay. And you'll say, oh, yes, of course. Well, of course, they're on Biscayne Bay. Maybe that'll be useful. So you write that bay is a useful rhyme. And you make a list of rhymes that are in some relevant to what you're trying to say and then you use them.

GROSS: The more you write, do you feel like you've used up rhymes?

SONDHEIM: Oh, yeah.

GROSS: Like you can't use a rhyme you've already used? A choice, a certain...

SONDHEIM: Well, that's certainly – that, now, that's certainly true of any kind of trick or...

GROSS: Give me an example of the kind of trick rhyme you're talking about.

SONDHEIM: Oh, goodness. I don't now. Soul stirring and bolstering in "Follies," you know. You use that once, you don't use it again. Lottie dottie and nobody. You don't use that more than once. Or if you do, you're a fool. And I probably have used them more than once but I don't think so.

(SOUNDBITE OF LAUGHTER)

SONDHEIM: So that's what I mean. Whereas, yeah, of course you're always going to end up rhyming day and may and say over and over and over again, you know, from song to song, show to show, because they're useful and they're words that have many meanings and many connotations. So that...

Transcripts are created on a rush deadline, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of Fresh Air interviews and reviews are the audio recordings of each segment.