First Species Counterpoint, Rule by Rule
Species counterpoint isolates one compositional decision at a time. First species isolates the hardest one: how to make two voices sound like two voices.
A cantus firmus is a tune with no accompaniment, eight to sixteen notes long, that moves mostly by step and comes to rest on its final. First species asks you to add exactly one note against each of those notes: two voices, both in whole notes, moving together through the whole exercise. Nothing about the task sounds difficult until the fourth or fifth note, when the obvious moves have run out and the remaining options are all wrong for reasons you have to learn to hear.
What makes the exercise worth doing is that it removes almost everything else. There is no rhythm to hide behind, no accompaniment to thicken the texture, no register change to reset the ear. Every vertical interval is exposed for a full bar, and the quality of the counterpoint is decided by which intervals you choose and how you travel between them.
What first species is actually testing
Species counterpoint reaches its canonical form in Johann Joseph Fux’s Gradus ad Parnassum (1725), which distilled the practice of sixteenth-century vocal polyphony into a graded method: note against note, then two notes against one, then four, then syncopated, then florid. Fux was teaching a style, not an abstract system, and the style had a specific sound: massed consonant intervals, each line shaped like a singable melody, cadences that come to rest on the final.
The order matters. In later species the added voice has rhythmic motion, which lets consonance and dissonance alternate quickly and lets melodic momentum cover weak moments. In first species everything is held. Two voices that are consonant but inert will sound inert, and two voices that are individually beautiful but collide will sound wrong immediately, and you will not be able to explain why until you look at the intervals.
So first species is really testing three separate skills at once: choosing consonant intervals, moving between them in a way that keeps the voices independent, and shaping each line as a melody. Keep them separate while you work. Most beginner errors come from trying to satisfy all three in one pass.
The intervallic framework
Consonances are divided into two classes, and the division carries real information. Imperfect consonances (thirds and sixths) can be used freely, including in parallel motion. Perfect consonances (fifths, octaves, unisons) are structurally stronger and therefore rationed: they open the piece, close it, and mark important points in between.
| Interval against the cantus | Status in first species | Conditions |
|---|---|---|
| Third, sixth (imperfect) | freely usable | the substance of the middle of the exercise |
| Fifth, octave (perfect) | usable, sparingly | approach by contrary or oblique motion |
| Unison | opening and close only | a mid-exercise unison erases the two-voice texture |
| Fourth | not usable | a dissonance in two-voice writing against the lower voice |
| Second, seventh, any altered interval | not usable | dissonant |
| Parallel fifths or octaves | forbidden | both voices move in the same direction by the same perfect interval |
| Hidden (direct) fifths or octaves | forbidden | similar motion into a perfect consonance |
The unison rule surprises people. A perfect unison is the most stable interval available, so it seems like a natural resting point. In two voices it is the opposite: at a unison the two lines have merged into one pitch, the texture momentarily has one voice instead of two, and the ear loses the independence that the whole exercise exists to demonstrate.
The rule against parallel perfect intervals has nothing to do with ugliness. Fifths and octaves are so stable that a pair of them in a row fuses the two voices into a single reinforced line; the independence that makes counterpoint counterpoint disappears. Fifths and octaves arrived at by similar motion produce a milder version of the same fusion, and in two-voice writing the damage is audible: the upper voice seems to slip into the lower voice’s track.
Contrary motion is the fix, and it is worth thinking of it as the engine of the style. When one voice rises while the other falls, both lines are audible as lines, and any perfect interval they arrive at sounds like a convergence rather than an accident. Oblique motion, where one voice holds while the other moves, is nearly as useful.
Play that pair of chords and listen to how quickly the interest drains out of it. Both voices step up, the fifth stays a fifth, and the result is not two voices but one thick one. Now the same bass with a contrary upper voice:
The interval changes from a fifth to a third, the upper line has its own direction, and the ear registers two parts. Nothing in the pitches changed except one note.
The added voice still has to be a melody
A counterpoint that satisfies every interval rule can still be a bad counterpoint, because the melodic constraints are doing half the work. The added voice should be singable by a person with an ordinary range, which means staying inside roughly a tenth and avoiding anything that would make a singer reposition.
The practical melodic rules are these. Stepwise motion should predominate; the exercise is not the place for a chain of leaps. Leaps are permitted up to the fifth and to the minor sixth, plus the octave, and the major sixth leap is generally avoided in strict two-voice writing. Any leap larger than a third should be recovered by step in the opposite direction, which is what makes a leap sound like an excursion rather than a series of disconnected pitches. Repeated notes are not part of the strict style at this species. Outlining a tritone, meaning the augmented fourth or its inversion, is forbidden, both by direct leap and by two leaps that trace the interval between them.
Then there is the shape rule. The added voice should have one high point, reached deliberately, usually by step, and not returned to. A line with two climaxes has no climax. A line whose highest note is also its first note has nowhere to go.
That example is legal almost everywhere and dull: a chain of parallel thirds, then a contrary arrival at the octave. Parallel thirds are permitted precisely because moving between imperfect consonances does not fuse the voices, but a whole exercise of them has no structural shape. Compare a version that spends its perfect intervals where they count, opens on the octave, uses thirds and sixths through the middle, and returns to the octave at the close.
The cadence deserves special attention because it is the only place where the rules relax in a specific direction. The final interval should be an octave or a unison, approached so that the finality is unmistakable, and in first species the standard close is a stepwise approach into the final in one or both voices. That is the ancestor of every cadence you will meet later in tonal music, and it is worth noticing here where it comes from: the counterpoint’s ending is a melodic event before it is a harmonic one.
Two passes, not one
The most common way to do this badly is to write the added voice note by note from left to right, choosing the best available interval at each step. That method has no view of the whole line, so it produces the equivalent of a walk that never leaves the block.
Write it in two passes instead. First pass: decide the skeleton. Place the opening interval, the closing interval, any perfect intervals you intend to use, and the climactic note of the added voice. Write those five or six decisions down. Second pass: fill the gaps with stepwise motion, favoring thirds and sixths, and let the melodic requirement that leaps recover by step do the work of keeping the line coherent.
Then check it the other way around. Read the added voice alone and ask whether it is a melody you would want to sing. Read the cantus alone, which you should already know is one. Then read the intervals from bottom to top, note by note, and name each one out loud. Errors in first species are almost never subtle; they survive precisely because the student stopped checking.
Where the exercise misleads people
The rules are a residue, not a foundation. They describe what sixteenth-century polyphony does most of the time, and Fux’s codification of that practice served a pedagogical purpose in a later era. Treating them as laws of music leads to the conclusion that parallel fifths are intrinsically wrong, which is false in most of the world’s music and a great deal of European music after 1890.
First species also teaches nothing about dissonance treatment, because there is no dissonance to treat. The interesting half of counterpoint begins in second species, where a passing dissonance on a weak beat is legal and the same note on a strong beat is not. If you do only first species, you have learned interval grammar without learning syntax.
And the exercise cannot measure quality, only legality. A counterpoint with no faults can be lifeless, and a slightly irregular one can be memorable. The rules exist to protect the independence of two lines; independence is what you are actually after.
How to work with it
Set yourself a cantus firmus, sing it until it is memorized, then write the counterpoint in the opposite direction. If you write below the cantus, the melodic shape you naturally favor will be wrong, and you will have to think about register. Sing each line separately before playing them together; the voice finds parallel fifths that the eye and the keyboard both forgive. Play the finished exercise on the virtual piano with the two voices an octave apart, which makes the intervals visible as well as audible.
Then invert the exercise. Rewrite the counterpoint below the cantus instead of above, keeping the pitch classes unchanged, and see which of your intervals survive the flip. Most first attempts do not survive it, and that failure is the entrance to the next subject.



