The four collections of smaller piano pieces, comprising all of Brahms’s latest piano works, were published in 1892 and 1893 as  Fantasien [‘Fantasias’] op.116, Drei Intermezzi [‘Three Intermezzos’] op.117,  Klavierstücke [‘Piano Pieces’] op.118, and  Klavierstücke op.119. The titles of individual pieces in these collections include Capriccio, Intermezzo, Ballade, Romanze [‘romance’], Rhapsodie [‘rhapsody’].
The Klavierstücke op.119 are: 1. Intermezzo, B minor, 2. Intermezzo, E minor, 3. Intermezzo, C major, 4. Rhapsodie, E♭ major. (The fact that Brahms originally intended to call his earlier B minor rhapsody, op.79, no.1, ‘Capriccio’ shows that he may have used such terms rather loosely. ‘Intermezzo’ can be seen as an umbrella term under which Brahms could collect anything which was neither capricious nor passionate.) These four pieces are Brahms’s last works for piano (The 51 Übungen für Klavier [‘Exercises for the Pianoforte’] were published in 1893 without opus number). He completed them during his summer holiday in Ischl, Upper Austria, in 1893, the first intermezzo being written in May and the following three pieces in June.
Since Brahms has combined these 18 character pieces in collections, he may have included some earlier compositions, and it is quite possible, although there is no definite proof, that some works—such as the E♭ major rhapsody—may have been conceived before 1892. Two earlier collections of smaller lyric piano pieces,  Klavierstücke op.76, and Zwei [‘Two’] Rhapsodien op.79, date from 1871-79 (published 1879 and 1880 respectively).
The poetic mood of the first intermezzo from op.119 belies its vague title. In a letter from May 1893, Brahms tells Clara Schumann:
"I am tempted to copy out a small piano piece for you, because I would like to know how you agree with it. It is teeming with dissonances! These may [well] be correct and [can] be explained—but maybe they won’t please your palate, and now I wished, they would be less correct, but more appetizing and agreeable to your taste. The little piece is exceptionally melancholic and ‘to be played very slowly’ is not an understatement. Every bar and every note must sound like a ritard[ando], as if one wanted to suck melancholy out of each and every one, lustily and with pleasure out of these very dissonances! Good Lord, this description will [surely] awaken your desire!
Brahms sounds somewhat ironic here, but Clara Schumann was enthusiastic and asked him to send the remaining pieces of his new opus.
The words ‘melancholy’ and ‘with pleasure’ aptly describe the decadent ‘Schoenbergian’ atmosphere evoked by those aimless opening harmonies. In fact no clear tonality can be perceived in these first three bars. The very first chord, for example, could be a B minor 7th chord superimposed on an E minor triad. The entire A section (bars 1-16) eludes the tonic and only the coda (bars 55-67) ends in B minor in a mood of deep resignation. The effect of wistful intimacy is not a result of improvisational looseness. The composer has calculated each effect in minute detail to create one of the most delicately wrought miniatures imaginable.
Brahms’s meticulosity in delineating the polyphonic texture, often produced by individual articulation in different voices, is almost beyond imagination and makes enormous demands on the sensitivity of the performer.
In the middle section of the piece (bars 17-46) the warmer D major is supported by more consonant harmonies, a less polyphonic texture and a lilting slow waltz rhythm. The motivic relationship can be observed in the top voice (compare bar 1 with bar 17).
To appreciate Brahms’s incredible expertise in manipulating the mood by slight changes in articulation and voice leading, compare bars 17-20, the opening of the middle section, with its repeated statement 31-34: In the later statement the blissful lilt has been ‘corrupted’ by an additional chromatic middle voice echoing the faded memory of the previous phrase (bars 29-30) and foreboding the immanent ecstasy of the climax (bar 39). Even the waltz rhythm in the left hand appears less comfortably declamated because the composer has renounced the use of legato. Only the somewhat decadently syncopated top voice maintains its articulation. Thus we have three voices—each with its individual articulation and rhythm.
The E minor intermezzo could almost be called ‘monothematic,’ but each time the theme appears in a completely different guise which makes it hardly recognizable.
Brahms has a very marked predilection for subtle rhythmic shifts. In the opening of the E minor intermezzo (Ex.14) the ¾ meter is disturbed by the sforzando and sostenuto in bar 2. In bar 3 a ‘correction’ is suggested, but already in bar 4 a much more serious displacement of the meter is occasioned with an apparent ‘wrong’ repeat of the theme on the third beat. The weak B minor sixth chord in bar 5 can no longer be perceived as a downbeat. The subsequent syncopated repetitions of the d♯ in the top voice completely distort the original feeling of triple meter. It is mainly through these rhythmic shifts that the music evokes a mood of nervous agitation and anxious searching.
But also the gracious playfulness of an apparent naïve melody, such as the main theme of the C major intermezzo, owes quite a lot to subtle rhythmic manipulations.
The freshness of C major is modified by presenting the melody as the middle voice, certainly lending it a more mellow hue. Brahms’s rhythmic manipulations involve the dichotomy of the first and fourth beats. The strong bass note C in the opening bar seems to establish a very certain downbeat, but this is contradicted by the right hand melody suggesting a downbeat on the fourth eighth note.
The length of the phrase is twelve bars, subdividing into two six-bar sections. The first six bars can certainly be heard as two three-bar units whereas the second six-bar section can rather be perceived as three times two bars. The second six-bar sub-phrase functions rhythmically as a giant Hemiola (2´3=>3´2).
This rhythmic gracefulness is opposed by the middle section of the piece. Two eight-bar phrases, subdividing into four-bar units, try to ‘correct’ the twelve-bar dream in a most energetic manner.
The smooth legato opening of the recapitulation (starting with bar 41) is hiding a lot of impatience causing the undulating Hemiolas of the first two bars to explode in a sudden outburst. As a result of this the main theme can hardly be perceived as the entrance of the recapitulation, and the underlying song form of the intermezzo becomes almost unrecognizable for the listener. Therefore it is arguable that this piece is in Binary form and the B section begins at bar 49 where new material appears.
Brahms’s experiments with rhythm and phrase lengths are also apparent in the E-flat major rhapsody which for 60 bars maintains 5 bar phrases. An insensitive performer will always present these first five bars as five times 2/4 rather than the implied two times 2/4, plus two times 3/4. More scrutiny in observation would reveal the absence of an accent on the fourth bar in connection with an inversion of the one fourth-note/two eighth-notes pattern in the left hand, clearly shifting the feeling of downbeat to the second beat of bar 4. The result is not a march rhythm (as described in many superficial analysis’) but rather an exalted heroic declamation.
The ‘grazioso’ second theme (starting bar 93) is constructed by eight-bar phrases that do not subdivide into four plus four, but rather into three plus two plus three. Together with the ‘lute’ like arpeggiated accompaniment, these uneven phrase lengths project a certain teasing charm.
The rhapsody from Op.119 has been criticized for its rather crude form and medieval austerity. But the form has to match its content and a complex polyphony or a sonata form like development would surely disturb the archaic character of this magnificently heroic epic. The fact that Brahms uses extremely complex polyphonic devices in the intermezzos, but refrains from employing them in the rhapsody, proves that his enormous technical expertise as a composer is applied to the service of the character his music is to convey.
Interestingly, the piece ends in E-flat minor, the parallel minor key to where it started (E-flat major). While it is not unusual to end a minor-key composition in the parallel major, it is much less common to find a work constructed in this manner.
Brahms’s achievements as a Romantic composer are not so much the discoveries of new Romantic traits but rather his ability to rely on his constructively Classical inclinations and enormous composition technique to bestow substantiality upon the Romantic language.