Arolas' “Unfinished”

(Zur deutschen Version)

The evolution of Eduardo Arola's Una noche de garufa has been described several times in the literature of tango. Francisco Canaro reported in his autobiography that in 1909, as he was performing with his group at a café, a young lad carrying a bandoneon entered with a group of friends. Canaro was told that the young man, Eduardo Arolas, had composed a tango and he asked to hear it. Arolas readily complied with the request and played Una noche de garufa, his first composition. At the time, Arolas did not know how to read and write music, and so Canaro and others offered to write it down for him.

After hearing Una noche de garufa, Canaro included it in his repertory. It became a successful tango. In 1911, a piano score was published and two years later, Arolas recorded Una noche de garufa with his own group. Thus appearing at the height of the tango fever that erupted in Europe and the Americas before World War I, the piece also became known and recorded outside of Argentina.

For a critical assessment of Arolas' work as a composer, Una noche de garufa provides an interesting point of departure—not only by being his first known composition, but also because it was conceived before Arolas dedicated himself seriously to the study of music. The piece shows how a young, aspiring musician composed a tango without actually knowing how such a piece was supposed to be made up. It is an example of composition “by ear”.

1. The Formal Structure of an Instrumental Tango

At the time when Arolas composed Una noche de garufa, most tangos were written as instrumental pieces, and most of them consisted of three sections (hereafter indicated as A, B, and C, respectively). The last of these sections (C) was usually entitled a “trio” in the scores. Each section was normally 16 measures long and structured into two musical phrases, antecedent and consequent, of eight measures.

The trio, section C, merits further mention. It was commonly set in such a way that it provided a conspicuous contrast to the other sections. The difference could manifest itself melodically, rhythmically, or harmonically. Furthermore, the trio commonly served as a sort of midpoint after which one or both of the previous sections were repeated again. A performance sequence of AABBACCA or AABBACCAB would have been typical. The exact sequence could vary; there existed no iron-clad rule that a repetition scheme had to be executed in a given order. The order of play was ultimately left to the performers. But by convention, a return of section A and/or B after C, the trio, was common and expected.

1.a The Tonal Structure of an Instrumental Tango

Parallel to the formal structure ran the tonal organization. Usually, at least one of the sections was set in a key different to the main one of the piece. If the main key was in major, then the contrasting key was conceivably the dominant and/or subdominant. If the main key was minor, the contrasting key could be the parallel or relative major one.

Corresponding to the repetition scheme, there existed no firm rule how the tonal contrast was realized, but certain conventions predominated. Repetition scheme and tonal organization corresponded also on another level: the last section played usually concluded in the main key of the piece. That is to say, if section A was in the main key, then the repetition scheme could have been ABACA, for example. Likewise, if section B was in the main key, the repetition scheme might have been ABCAB. Again, there existed no fixed rule on how this was executed, but certain conventions prevailed.

2. Arola's Una noche de garufa

2.a The Formal Structure

As a musical form, tangos are simple structures, usually consisting of two or three sections of 16 measures length. Una noche de garufa does not follow this pattern. Each section consists of a different number of measures:

A: 8 measures

B: 18 measures

C: 16 measures

Only section C appears to correspond to the conventional formal structure. Section B, being 2 measures longer, deviates from the common pattern only to a small degree. In fact, we shall see below that the two extra measures are a consequence of the particular tonal development in this section.

With only 8 measures, section A is, however, strikingly different. We mentioned above that 16-measure sections were customarily divided into two phrases, antecedent and consequent, of 8 measures each. Hence, it appears that section A is truncated and lacks the consequent phrase. This does not necessarily mean that the section is flawed, but it is certainly odd and affects the formal symmetry of the piece.

2.b The Tonal Structure

Una noche de garufa carries the key signature of B-flat major. Section A begins in B-flat major, on the tonic (indicated as I in the following examples), and ends with a cadence on the dominant, an F-major chord (indicated as V in the examples).

Una noche de garufa, Section A

We have noted above that section A appeared to be “incomplete”, inasmuch as it consists only of an eight-measure antecedent phrase and is lacking the consequent phrase. This notion is corroborated by its harmonic development. The phrase ends on the dominant (V), which in music theory is called a “half cadence”. It is not uncommon for antecedent phrases to end with a half cadence, but one would expect a consequent phrase leading back to the tonic. This is, however, not the case and thus section A leaves an impression of incompleteness.

Una noche de garufa, Section A, Sexteto Carlos Di Sarli

The conventional two-part division into antecedent and consequent phrases of 8 measures is clearly discernible in section B. There is an anomaly, however: between antecedent and consequent phrases (measures 1-8 and 11-18 in the following example, respectively), two extra measures have been interpolated (measures 9-10).

The harmonic structure of the section is quite simple. The antecedent phrase consists of a series of alternating F and B-flat major chords whereas the consequent phrase shows a series of D major and G minor chords. The two interpolated measures anticipate the harmonic progression of the consequent phrase.

Both the antecedent and consequent phrases consist of a series of two alternating chords that are a fifth apart. Hence, they can be interpreted as a series of dominant-tonic progressions (indicated as V and I in the following example) . The chords of the antecedent phrase (F major and Bb major) are, in fact, the dominant and tonic of the main key of the piece, B-flat major. Due to this dominant-tonic progression, the resolution to the (incomplete) half-cadence ending of section A occurs here in the antecedent phrase of section B.

The dominant-tonic progressions in the consequent phrase of section B involve the chords D major and G minor. The phrase actually ends with a g-minor chord and does not return to B-flat major. This suggests that the consequent phrase is not in B-flat major as the antecedent phrase, but in G minor.

Here the two measures interpolated between the antecedent and consequent phrase (measure 9 and 10 in the example below) come into play. We have noted that they anticipate the harmonies that unfold in the consequent phrase. It is in these two measures that a switch from B-flat major to G minor takes place. In music theory, this is called a modulation, that is, a gradual progression from one key to another.

Una noche de garufa, Section B, Sexteto Carlos Di Sarli

Una noche de garufa, Section B, Sexteto Carlos Di Sarli

There is a correspondence between sections A and B: both begin in B-flat major, yet do not end on the tonic of that key but on a related sonority. Section A concludes with a half cadence on the dominant, and section B modulates to G minor, the relative minor key. Hence, both sections are tonally not “closed” but imply continuation.

Section C, by contrast to the preceding sections, is an example of regularity. It consists of 16 measures, divided into antecedent and consequent phrases of 8 measures. Both phrases begin and end with B-flat major sonorities and never venture into tonal areas other than the tonic. Section C begins and ends in B-flat major and is thus tonally “closed”.

Una noche de garufa, Section C

Una noche de garufa, Section C, Sexteto Carlos Di Sarli

3. The Repetitions

It was pointed out above that three-part forms like Una noche de garufa were rarely performed from the first to the last section in the order they were printed in a score. There was always some kind of repetition involved, but no fixed rule on how the repetitions had to be executed. Una noche de garufa is a tango that has been recorded many times, and a look at a selection of these recordings illustrates the point.

The list is by no means complete and the selection of orchestras is arbitrary. Included are 9 recordings made over a period of 40 years. The earliest one may be Arolas' own recording of 1913; the last one was recorded 1953 by Francisco Canaro and his Quinteto Pirincho.

To the right of the orchestras' names, the sequence in which the sections are played and repeated is indicated. Eduardo Arolas, for example, performed section A to C in order and then returned to A. Each section was played twice, thus resulting in a sequence of AA BB CC AA for the complete recording.

Common to all recordings is that they first play one cycle of sections A through C in order. Within this first cycle, A is always repeated but sections B and C are not repeated by all groups. Among the nine recordings, three different patterns of repetition exist—AABC being the most common one.

After the first cycle, all but one recording continue to repeat one or more sections. The second cycle shows more variety in the repetition schemes. Among the nine recordings, six different patterns can be identified—the recording of the Rondalla del Gaucho Relámpago being the simplest with no repetition at all, and Canaro being the most complex by repeating section A in a third cycle.

Two conclusions can be drawn from the above table. It is striking that only two out of nine versions (Adolfo Pérez' and Ricardo Tanturi's) use the same repetition scheme. While an adherence to certain conventions is discernible in all versions, the orchestras appear to have set value in recording their own versions. All versions begin with section A. However, all three sections were considered suitable as an ending. (Since the selection of recordings was made randomly, it should not be inferred that the distribution of endings presented in the table is statistically correct.)

4. The Endings

If one asks musicians how the tonality of a composition is determined, one will likely be told to look at the chord on which the piece ends. This presents a quandary with the piece under discussion since all sections were used as a conclusion. Yet, each section ends on a different chord: F major in section A, G minor in section B, and B-flat major in section C.

We have noted that section A ends with a half cadence, thus, on the dominant of E-flat major. If the recordings ending with section A are considered to be in Eb-flat major, then their ending is “unfinished”. The dominant, the final sonority, is never resolved. Interestingly, this is the ending chosen by Arolas.

Given the extent of Arolas' musical training when he composed the piece, it is unlikely that he consciously chose to end section A with a half cadence in order to write a tango with an “open”, unresolved conclusion. It probably just sounded interesting to him to have the resolution to the tonic follow in section B. A repetition of section A at the end was, however, expected and added during performance in order to comply with the convention.  

Una noche de garufa, Orquesta Eduardo Arolas

From its early beginnings, the music recording industry was an international business. As a consequence, music from all parts of the world was recorded and distributed even in the remotest corners of the world. If a recording company did not have production facilities in a given country, either an engineer would be sent out to make recording on site, or the music would be recorded in their facilities at home. It is, therefore, not surprising to find a military band from New York and a “gypsy-orchestra” from Paris among the groups that recorded Una noche de garufa.

For both these orchestras, tango was a novelty. It is highly unlikely that they were familiar with the performance practice of tango in Buenos Aires. (They surely had not heard Arolas perform, whereas Pacho Maglio or Canaro certainly did.) As a consequence, they had to make sense of the printed score without an extensive experience in playing or listening to tangos.

It is interesting to see, then, that both orchestras opted for a tonally closed interpretation by playing two cycles of the three sections and concluding with section C, the trio. To close a piece with the trio is unconventional inasmuch as this section typically presents a contrast to the other sections. It denotes a point of departure rather than one of completion, as one would expect from a conclusion. In this case, however, it provides tonal closure in B-flat major: the piece ends in the key in which it started out.

Una noche de garufa, Military Band, New York

The version of Una noche de garufa best known by tango dancers today, the recording by Ricardo Tanturi, pursues the third possible conclusion and ends the piece with section B. This has certain advantages over the other two choices: the piece does not conclude “incompletely” with a half cadence, and it does not end with the contrasting section. Bringing the piece to a close with section B provides an ending with a full cadence and highlights the contrasting nature of section C by surrounding it with sections A and B (AB C AB). This version, then, is symmetric and formally closed. It is not closed tonally since it ends in G minor, but it ends at least in a key related to E-flat major and terminates with a full cadence.

Una noche de garufa, Orquesta Típica Ricardo Tanturi

Our discussion of the different versions suggests that each one represents a possible solution to aspects that from a music-theoretical perspective are problematic. None provides an answer to all the questions that were raised by our analysis. This incongruity, we submit, is due to Arolas' lack of experience as a composer . Later works by Arolas—Cardos, for example—present a very different picture. These pieces are carefully planned. The sections are proportional and harmonies are meticulously balanced and used with great effectiveness. But between these two extremes in Arolas' work lie years of study and experience.

Tango musicians, whether they were composers or performers, have left no records on why they decided to compose or play a piece one way or another. Shop talk was not something to be related to the public. We do not know why Pacho Maglio ended Una noche de garufa with section A and Di Sarli continued to section B. We can only hypothesize.

We also do not know if Arolas would have composed his tango differently after his musical studies. However, we do have evidence that he became aware of musical conventions and possible criticism for irregularities. His “tango characterístico” Anatomía (written before 1917) contains a footnote regarding an anomaly in the first section of the piece, which is one measure shorter than the conventional length:

NOTE = I ask the teachers who perform this tango to excuse the 15 measures in the first part, but it is done deliberately.

Arolas' note is remarkable in various respects. It addresses an issue we have noted in Una noche de garufa as well, namely, an asymmetry in the phrase structure. However, it is, to our knowledge, a unique occurrence that a tango composer adds a justification for an irregularity to a score. Pointing out that it was introduced intentionally prevents, of course, being accused of having made a mistake. We may also assume from which party the composer expected criticism to arise. It is not fellow musicians like Canaro or Maglio he is addressing, but the “teachers”. Arolas had learned his lessons.

Eduardo Arolas  (≈ 1910)

© 2018 Wolfgang Freis