Aerosmith - I Don't Want to Miss a Thing (from "Armageddon")

Music by Diane Warren

Piano Arrangement by Tom Roed from musicnotes.com was used for analysis.

FORM

INTRO-ABABCBB-CODA (Verse = A, Chorus = B, Bridge = C)

KEY

The song is in D major, but the introduction is in E mixolydian.

INTRODUCTION

As I said earlier, this eight bar introduction is modal, not tonal. There is no dominant harmonies and no leading tone C# that establishes D as a tonal center. Instead, we have a bass movement resolving to E. Also, all pitches of E mixolydian are present -- E, F#, G, A, B, C#, D [b7]. Chords on the 2nd, 4th, 6th, and 8th bar are interesting because A, E, and B are sustained. A/E, E/B are perfect fifths from each other and create what is called a quintal harmony.

SECTION A (Verses)

There are a couple of odd things going on here. The ii7 (Em7), a predominant harmony, usually moves to the dominant. In this case it moves back to I. Also the 4-3 suspension in the Asus chord (V4) doesn't resolve. But a kind of cool spot is when the ii7 moves to iii7, creating a ii7-iii7-IV7 parallel chord progression, also described as harmonic parallelism. A unique feature about the bass throughout this song is it's stepwise movement. Using inversions of chords allows the bass to move minimally, creating a smooth line.

CHORUS 

Here, again, a ii7 interrupts the resolution of the V6 chord, but also creates a kind of deceptive-cadence-like effect. Em7 contains B and D that are also in a vi chord which is usually used in a deceptive cadence. The same stepwise bass movement (except the leap to E) is seen here as well.

BRIDGE

At first, just like in the Beatles' You're gonna lose that girl, I thought there was a tonicization of bVII in the bridge. It sounded as if we modulated to C major, but I was confused because there were no strong cadences establishing C major, and no pivot chords preparing the modulation. Below is my initial Roman numeral analysis. 

I soon realized that labeling the Dm7 as a ii7 didn't make sense. Dm7 is ii7 in the key of C major but i7 in D major. I've never seen a i7 chord act as a pivot. Also, while it would not be surprising since we saw this before, the V6 chords do not resolve properly. Below is my second Roman numeral analysis where I used secondary chords instead of showing a tonicization.

While this creates a III6 chord, I think it makes much more sense. The III6 is a result of a descending fifths sequence, explaining the odd label (diagram below, I changed first inversion chords to root position for clarification). Following the descending fifths is an ascending fifths (also notated below).