| home | authors | calendar colophon | links | newsgroups | newsfeed | new | printer version |  
volume 18
september 2015

Plagiarism or inspiration?


  On the relevance of melody as a marker for plagiarism in Pop and Rock music
by D. Pinter
  Some songs closely resemble one another. But, when can we say that the similarities are coincidental, come down to plagiarism or to something in between that may be labelled "inspiration"? Adding about a dozen songs to some well-known cases, D. Pinter here unfolds a critical approach to judge alleged plagiarism in popular music.

Plagiarism or inspiration? Musical plagiarism nowadays is a hot topic, and it seems that it will remain so for a while. Alleged cases come easily to the mind, ranging from recent songs like "All About That Bass" (2014), "Blurred Lines" (2013) and "Get Lucky" (2013) to older rock anthems like "Sweet Child Of Mine" (1987), "Hotel California" (1976) and "Stairway To Heaven" (1971). And, that is just to name a few of the song titles that over the last decades came up in the headlines. The verdict is not clear for all cases and we are facing the question: where lies the threshold of musical plagiarism?

  The US copyright law vaguely writes of concepts like "idea" and "expression", and only hints at the way in which these terms can be translated into the language of music. The closest description given of plagiarism, is the existence of a "substantial similarity" between two works. However, precise, quantitative algorithmic rules are not specified to determine what exactly constitutes a "substantial similarity". No wonder, the resemblance between songs depends on so many different factors and their combinations that it's near impossible to integrate all of those in one ultimate rule of thumb. Since the copyright law specifies no rules, any expert witness can consider any similarity as substantial or just as easily deject it as superficial. The verdict generally relies on the common practice of prior cases, applying various methods and tests to ease the decision. In this study I will try to take a step further by specifying some rules and providing some examples supporting them.
  In court, the question of plagiarism in the end always results in a "yes or no" decision. From a musicologist point of view plagiarism sometimes is a question of probability. For every clear example of plagiarism there are numerous uncertain cases filling the large grey zone we may call "inspiration". A major point of this article is that there are cases where inspiration is undeniable at play, and sometimes even admitted, but where the music itself is still far from being a pirated design.
  Musical plagiarism — as proposed here — is primarily a matter of a close similarity of a rather long (two bars or more) melody or some fragments of melodies reinforced by arrangement-related elements. Inspiration concerns more subtle melodic similarities and all of the accompaniment, chord progressions, special effects, style of arrangement, lyrical themes, and so on. In short, without any very close similarity of a longer melodic fragment we usually cannot talk of plagiarism. In this article I will propound why melody is so important when looking for plagiarisms and why in many cases complementary evidence of arrangement-related similarities will be necessary for a final judgment. I will substantiate my case with a fair number of cases.
2 Levels of similarity. A similarity between two songs or, in a broader sense, two musical pieces can be registered on various levels, such as melody, style, harmonic progression, structure, phrasing, rhythm, not to mention a number of things that are not directly related to the music itself: the lyrics, the arrangement, the video clip, and so on. For now we will focus our attention on the music-related factors.

Harmony. Harmony is a primary element of Pop and Rock music. However, it's usually a secondary factor to consider in cases of plagiarism. A generally accepted rule — though not included in the letter of law [1] — is that chord progressions are not copyright-protected; they are free musical panels commonly accessible to everyone. This status primarily adheres to common chord progressions, a category to which the vast majority of chord progressions in the world of Pop and Rock music belongs. Many of these progressions have been in use for ages, like the turn-around: I —> vi —> IV —> V. [2] This cliché, widely used in doo-wop songs around 1960, for instance also opens the baroque originated "God Save The King".

What, then, can we say about more unconventional chord progressions? The Eagles' well-known hit-song "Hotel California" (1976) is based on an easily recognizable "signature" progression of eight chords — not a quite frequently used one. A rare example is "Love Theme" (1983) from the Flashdance soundtrack. But, Jethro Tull used the same progression already back in 1969 in their song "We Used To Know". Is this a case of "great minds thinking alike", or a matter of influence? We don't know, but the mere chord progression itself was never subject of debate. To date this curious chord progression could not be traced back earlier than 1969, though chord progressions with a "substantially similar" chromatic descent built into the harmony are known. [3]
The lack of earlier examples by itself is a not a very convincing sign of originality, since a music analyst cannot browse through more than just a few thousand songs to find an even earlier example. Rather special, but only four chords long, is the progression I —> II —> IV —> I with a chromatic inner line, [4] which so far could not be found prior to the Beatles' song "Eight Days A Week" (1964). Can this progression be considered to be invented by the Beatles? If so, can those who followed their example be sued for using it? For the time being it's a cliché that belongs to the public domain. Procol Harum's song "Homburg" (1967) was one of the very first to follow the lead and Cee Lo Green more recently scored a major hit, "Forget You" (2010), also featuring this chord progression.
  Similar or identical chord progressions are often referred to in the context of plagiarism. The importance of chord progressions, however, usually comes second. This means that chord progressions by themselves are not fundamental to questions of plagiarism. On the other hand, they can strongly enhance our sense of similarity caused by other factors. This rule works for common chord progressions to a lower and for special and rare chord progressions to a higher degree. In this respect the harmonic rhythm also counts. Evenly timed chord changes are too common to be treated as a relevant factor of plagiarism. A varying harmonic rhythm, on the other hand, is already a much more characteristic detail.

Rhythm. Rhythm is another important mark of Pop and Rock. But, like common chords progressions, generic rhythmical patterns cannot be copyright-protected. And, like more specific chord sequences, special rhythmical patterns are not protected either. We don't know about any cases of plagiarism focusing primarily on rhythm. Rhythmical similarity is not even an easy-to-notice phenomenon unless you are focusing on it.

Next to the overall rhythm, melodies have their own rhythmical dynamics. Comparing melodies, a perfect matching presupposes both identical pitch and identical rhythm (timing). Once the pattern or a single note is shifted backward or forward in time, the matching is over. You can test this phenomenon on the intro riff of The Kinks' song "You Really Got Me" (1964), which you can try to interpret in two different timings (Table 1):

      4 . 1 . 2 . 3 . 4 : beats
          7,1 1 7,1     : "false" timing as a first-ever
                          listener experiences (before
                          drums enter). 
    or: 7,1 1 7,1       : real timing
  Table 1: "False" and real timings in "You Really Got Me"
[In this and the following tables numbers refer to scale degrees (here in minor), while commas and apostrophes denote lower and higher octaves respectively]
  Though the notes are the same in both samples, the timing varies and the riffs also do sound substantially different due to the shifted placements of the accents. What is left is just a subtle feeling of similarity. On the other hand there are cases where an imperfect match in timing can, to a certain extent, be counted as contributing to a case of plagiarism. In the case of an "isolated" note — with no notes falling on the neighbouring eighths before and after the note's impulse — a rhythmical lagging can keep the feel of similarity alive. See for example the case of "Get Lucky" vs. "Love Life" below. Sometimes a shifted group of identical notes can also be assigned a lower weight. See the example of "Come As You Are" vs. "Eighties" example or of "My Sweet Lord" vs. "He's So Fine" below.

Sound. "Sound" is a concept that designates the chemistry between instruments, arrangements, the voice and style of the singer, which includes many subtle details and special effects. In Pop and Rock music the concept is used to differentiate between genres and to designate the unique qualities of performers. It is generally accepted, though, that like chord progressions and rhythmical patterns a specific sound cannot be patented.

A provoking question in this respect is: is it legal to make a close copy of somebody else's song's backing track (with minimal melodic content), and then add a self-written melody over that? I'll keep this question open because the answer depends much on the uniqueness of the particular accompaniment.

  Some songs sound deceptively similar in style to other artists or bands. Bruno Mars's "Locked Out Of Heaven" (2012) has a distinct Police sound in the verse, while his other hit song "When I Was Your Man" (2013) has a certain Elton John sound. One can even name the particular inspiring songs: respectively "Message In A Bottle" (Police; 1979) and "Sorry Is The Hardest Word" (Elton John; 1976). There are subtle details in the music that create this impression.
  The rhythmical hook of "We Will Rock You" (Queen; 1977) is unique, primarily for its distinct sound, while its rhythm is as simple as can be. It consists of only three notes without pitches in one bar. Is it public domain from day one? Some songs that were clearly inspired by this hook feature drums instead of footstomps and handclaps like "E.T." (2010) by Kate Perry. Even though anybody can tell it's been taken from "We Will Rock You", this is still not a case of plagiarism. The choice of footstomps and handclaps over an identical rhythm would be a threshold case, since this characteristic sound can be seen as a distinct mark of the whole song that is as important as the melody itself.
  Many genres and subgenres have emerged from sound-alike (and look-alike) phenomena, following the style of a certain successful band or song. The aesthetic models of subgenres later on changed, depending on actual trends. Anyone claiming that "you stole my sound" should point out the ingredient of his/her/their signature sound invented exclusively by him/her/them. And even if some of that is provided, we only have proof of direct inspiration. In the case of "Come As You Are" (Nirvana; 1992) we will see, that a closely similar guitar tone also can strongly enhance one's feel of similarity. The auto-tune effect used on vocals is known for making songs resemble Cher's "Believe" (1998), even though this song was very far from introducing this effect into the field of Pop and Rock music.

Melody. After having discarded harmony, rhythm and sound, at least as prime markers of plagiarisms, we are left with the melody. Even though not emphasized in the law, melody usually is the deciding element in cases of plagiarism. Melody is an ordered collection of pitches and as such has shape as well as rhythm. Only a simultaneous matching of both these properties will count in full weight when deciding a case of plagiarism. Rhythm or melodic shape alone is only good for creating slight resemblances and may indicate inspiration. Similar but not identical fragments can instigate and amplify a sense of similarity. Especially lead vocals and hook melodies are the key subjects of plagiarism, while the arrangement of the backing track (including short solo fills, backing vocals) is more closely related to inspiration. This is also just a "generally accepted rule", and the proposed approach of this study.

  Interpreting the concept of "substantial similarity" is always a key step. The length of similar or identical fragments is definitely an important factor. The probability of an accidental matching will drop rapidly by every single identical consecutive note or chord. The first four notes of Beethoven's "Fifth Symphony", for example, make for an instantly recognizable piece of music. Its long closing note makes the sequence last for two bars. But, when we simplify the pattern to "impulses" of four pitches in one bar, the motif is not that unique anymore. The first two notes of the verse of Queen' "Bohemian Rhapsody" ("mama ...") are also instantly recognizable, but if we subsequently remove the backing track, the lyrics and the slide-up ornament, the fragment of two notes loses its recognizing character. Where is the limit? Is there a limit at all?
An unwritten rule says that five to nine identical notes in a row (or two bars) constitute the minimum threshold of plagiarism. Even though this rule is just an "urban legend", there is some truth in it, since the chance of accidental melody coincidence decreases "exponentially" by every added note, [5] and reaches a reasonably low level around eight consecutive notes or two bars. As we will see, there are melodic and rhythmical patterns where a series of eight consecutive notes is still not sufficient to label a song as plagiarism. Vice versa, there are cases where a 15/20 or 75% ratio of identical matchings can be sufficient, even if the longest identical fragment doesn't exceed the length of six notes in a row. And, in some cases only 4 to 5 corresponding notes with a sufficiently unique melodic shape and rhythm and similar lyrics can be enough to establish a case of substantial similarity. Below we will also offer some examples where a low rate of perfect matching notes can be (and should be) weighted by other close matchings as well.
  There are many common melodic patterns: scalar, pentatonic scalar, oscillating (horizontal, descending, ascending), recitative (horizontal), triadic leaps. Holding on to one note and stepping up or down one degree is the most common progression in sung melodies. To move in thirds comes out at a somewhat lower percentage, but not by much. Leaps over thirds (lead vocal) have got far more characteristic power. In instrumental melodies leaps over thirds are varying over a wider range. Two full bars of a constantly pitched tremolo of 16th notes can be a less characteristic coincidence than three notes moving in leaps over thirds.
There is a simple way of comparing melodies known as "the comparative method". [6] Going by the sheet music of two samples A and B, you simply count the amount of notes in the A sample that coincide with those in the B sample while starting at exactly the same point — negating the duration or legato/staccato articulation of notes. Differences in the home key and tempo are (to a certain extent) also set aside; they can be transposed and unified for the purpose of comparison. As a key indicator we can consider the length of the longest matching (or almost matching) sequence. And, of course, we have to check the ratio of matching notes compared to the sample size (the number of notes within the supposedly similar part of music).
  This test is easy to perform; we can observe the similarities and differences in a visual, quantized way. This method is useful, but sometimes misleading and therefore not fit as a sole criterion for testing plagiarisms. Simple generic melodies may result in a high ratio and long matching fragments, and still not be as relevant as a special melodic matching that is much shorter. Melodies may incorporate emphasized notes as well as "filler" notes and the former, no doubt, by far outweigh the latter.

Probability. We have already said that in most cases plagiarism is a matter of probability. That is to say that in and between two particular songs resemblances may be due to chance. These may or may not indicate undeniable inspiration or even plagiarism. The songs may share the same chord progression in the same harmonic rhythm, the same remarkable guitar fill, the same special rhythm pattern or a matching melodic fragment. We can assign probability values to each of these common details. We can multiply all the individual probability values, and so calculate a rough value of probability of accidental matching, let's say 1:50.000. We then may find new common details that further decrease this value.

  Now the question: for how many songs should we project this probability? A few 10.000 of songs that a particular songwriter has ever accessed; most of which belong to the category of well-known songs. Or do we have to project it to a million songs from all around the world and from all the decades of rock 'n' roll; most of which is obscure. This compares to winning a lottery with 30.000 or with 1.000.000 tickets filled. Clearly, much depends on a song's renown. When the source is a well-known song, one cannot defend one's case by saying that one has never heard it before. But if the source is an obscure song from an obscure band from an obscure country, then it increases the probability of accidental coincidence, or a common source of inspiration.
4 Cases and consequences. In this section we will browse through some songs, trying to locate signs of inspiration or plagiarism. We will see that melodic similarities by themselves often are not sufficient to denote a case as plagiarism and that corroborative evidence is needed. The examples will also show that in many cases next to the alleged original song one can find others with similar melodies. From this we can draw the conclusion that many supposed cases of plagiarism are in fact contingencies: a happenstance of songwriters unwittingly creating the same generic melody. We will start our musical exploration with a few examples of songs with a similar harmonic or rhythmical pattern and a case that shows that appearances may easily be deceptive.

First impressions may deceive. "Because" (1969) is a well-known Beatles' song written by John Lennon. In an interview Lennon once said that the song was inspired by Yoko Ono playing Beethoven's "Moonlight Sonata" but, on Lennon's instigation, played backwards. A thorough analysis verifies Lennon's claim, [7] but things are not as simple as the interview suggests. There are some special chord changes (even sequences) that appear both in Beethoven's piece and in "Because" — indeed in reversed order. And, Lennon probably would never have come up with these progressions without having listened to those chord sequences, and the song would have been significantly different without them. The melodic lines, though, have nothing in common with those of the "Moonlight Sonata".

  All in all, the source of the song may be unquestionably clear for a musical analyst. But, Lennon worked at it in a subtle and clever way. An audience test group surely could not hear the subtle similarities. One may wonder if even Beethoven himself would have recognized his own composition in the resulting Beatles' song, even in reversed form. It doesn't matter much: Beethoven died in 1827 and after 70 years music copyright is statute barred. We may call this song an example of inspiration, but not of plagiarism.
  Now let's look at an example of two songs with a similar rhythmical pattern. The song "Crazy Music" (1980) by the French disco duo Ottawan has a hook with a special "sharply" syncopated edge at its end (Table 2):
    1 . 2 . 3 . 4 : beats
    *   *   *  *
  Table 2: Rhythmical pattern of "Crazy Music" and "Get Up (I Feel Like Being A) Sex Machine"
  This rhythm is special and not too much other examples of it can be brought up. "Get Up (I Feel Like Being A) Sex Machine" (1970) by James Brown is one of the few ones that uses the same pattern with exactly the same notes: 1 —> 1 —> 2 —> 1 (degrees). It's a combination of a very basic melodic pattern with a special rare rhythm. Even though the matching ratio is 4/4, the four notes, even repeated, are still not enough to consider this hook plagiarism. It is very probable, though, that is a case of inspiration.
  The fact that first impressions may be deceiving is shown by a well-known Guns 'n' Roses' song: "Sweet Child O' Mine" (1988). Some two or three decades after its release, it was noticed that some parts of this classic sounds remarkably alike to "Unpublished Critics" (1981), a song released by the group Australian Crawl. The following similarities were pointed out:
  • the melody of the verse;
  • the chord progression of verse and chorus;
  • the guitar link/solo;
  • the chorus melody.
  The melody of the verses shares 8 identical notes out of the 32 notes of the "Unpublished Critics" sample, or 25%. This is not a rather high value, especially given the narrow range of the melodic line, ranging from 1st to 5th degrees — which, by the way, is a point of weak similarity. Four of the matchings are located in the first phrase, close to each other. And indeed, it's exactly these 3 to 4 matching notes that are responsible for one's initial impression of similarity. Without these notes the whole issue of plagiarism would probably never have popped up.
  Still, the matching ratio in the first phrase is below 50%. In the second phrase only the first note is identical, which is a key note here. The remaining part is very different. The overall rhythmical design of the verse is very different (18/32 or 56%) (Table 3):
      Bar 1-4
    4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 : beats
      55 5  4345 3  5 4314 31 45 4  4 : "Sweet Child O' Mine"
    1   55554343 1 5344 43 212        : "Unpublished Critics"
         *  ***       *     *         : matching notes

      Bar 5-8
      1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 : beats
      4 434314  4345 4 2              : "Sweet Child O' Mine"
          444441    115 3221          : "Unpublished Critics"
          *  *                        : matching notes
  Table 3: "Sweet Child O' Mine" vs. "Unpublished Critics": verse
  The chord progressions of the verses are based on the same standard progression: I —> bVII —> IV —> I, changing within the length of two bars. This progression, however, is a real cliché. Again, this coincidence contributes a great deal to our sense of similarity. The chords of the chorus are different, but similar, and the harmonic rhythm is identical. The chorus of "Unpublished Critics" uses the V —> IV —> I progression, while "Sweet Child O' Mine" goes with: V —> bVII —> I. The middle chords are different but interchangeable in the musical idiom of Pop and Rock, and therefore further enhance the subjective feeling of similarity.
  In both songs the guitar hooks share the same patterns: stepping one step downwards in 3+5 timing even though the two samples have not a single note in common. Moreover, "Unpublished Critics" incorporates yet another guitar fill that also clearly resonates with "Sweet Child O' Mine" as a bass figure. These similarities alone, however, are just details, and only fit to support a suspicion of inspiration. The chorus' melody is just remotely reminiscent: only the opening stretched-out "oh/ah" syllables and the subsequent stepping down are shared. The upper harmony vocal line in "Unpublished Critics" shows a closer similarity with the 3 —> 4 —> 2 melodic progression that sounds similar to what we can hear in "Sweet Child O' Mine". However, only the first note matches perfectly (Table 4):
    1 . 2 . 3 . 4 . 1 . 2 . 3  : beats
    3             4 2...       : upper harmony vocal
                                 in "Unpublished Critics"
    3           4 3 3 2 2...   : chorus of "Sweet Child O' Mine"
    *                          : matching notes
  Table 4: "Sweet Child O' Mine" vs. "Unpublished Critics": upper harmony vocal line
  Weighing all the details, we can judge this case a likely matter of inspiration, but not of plagiarism. The details of the guitar hook that — even without a single perfect match — also evoke reminiscences of "Unpublished Critics" and reinforce the suspicion of inspiration. It remains to be seen whether the borrowing was intentional or not.
  Moreover, there are more songs that resemble "Unpublished Critics". Listen for instance to the intro of "So Far Away" (1985) by Dire Straits. At first hearing, the lay listener may hold the muted guitar fifth chords in 3+5 rhythm for another "striking" similarity with "Unpublished Critics". Even the chord progression does sound the same, even though the chord functions and harmonic rhythm are identical with the chorus instead of the intro of "Unpublished Critics". An audience-test group could be easily convinced that it's a rip-off. Music experts, though, will be quick to see through these superficialities because they know that this type of intro is generic and appears in dozens of songs — a good example here is Bryan Adams' "Summer Of '69" (1984).

Coincidences. Similarities between songs may be insubstantial, or just due to serendipity. "Blurred Lines" by Robin Thicke and Pharrell Williams was released in 2013. Two years later, the song made the headlines as a case of plagiarism. The initial claims were talking about the overall sound being similar to "Got To Give It Up", a song released by Marvin Gaye in 1977. In a preliminary analysis the defendant's music expert, Judith Finell, brought up eight details of similarity. [8] Five of those details related to the melodies, one to the melody of the bassline and two to the arrangement. For the trial some of the details were updated with additional points brought in by a second musicologist.

  The melody-related points of the preliminary analysis, for the main part, were focusing on "substantial similarities" instead of identical melodic fragments. No wonder, among the included excerpts one cannot even find two [!] consecutive notes of melody matching in identic timing. In this case this trial is unique, because subtle signs of inspiration were presented as "substantial similarities" then blown up to direct one-to-one copies. No wonder the music industry in general gave negative feedback to the verdict of the first-round trial.
  Now check another example with a long melodic coincidence to be found in the lead vocal's melody. There is a descending melisma closing the first verse of "Blurred Lines" (3= b3 degree; modal). Compare this one with a melisma to be found in "Another One Bites The Dust" (Queen; 1980) and sung right after the interlude (Table 5):
    1   .   2   .   3   .   4    : beats
    5   4 3 4   3 1 3   1 7 1    : "Blurred Lines"
                                   (3 = minor 3rd degree)
    4 5 4 3 4   3 1 3   1 7 1... : "Another One Bites The Dust"
                                   (3 = minor 3rd degree)
        * * *   * * *   * * *    : matching notes
  Table 5: "Blurred Lines" vs. "Another One Bites The Dust"
  Here we have no less than nine [!] consecutive notes matching in identical timing. This is far more than just a local substantial similarity. Is it plagiarism? Not really, because the pentatonic melody combined with a simple rhythm and regular descending pattern is not unique enough. There must be far more similar examples in other R'n'B songs. Another detail to consider is that the melody is not repeated, and doesn't reach the total length that would turn even a section of the song substantially similar.
  Yet another example is the song "All About That Bass", the 2014 summer hit by Meghan Trainor. It copied, so was claimed, the chorus of the song "Happy Mode" (2006) by the Korean group Koyote. The case related to two full phrases, eight bars in length, which is a very long sample. The Koreans didn't sue, maybe because yet another song was found — "Contact" (1988) by Phish — with a similar melody and an even earlier date of release. Now, who did imitate whom? The length of the quote is far over the unwritten rule of eight notes or the two bar limit of plagiarism.
Can such a long overlap happen accidently? Yes, so it seems. Proof is given by a Hungarian school-camp song by the name of "Napkorong Az Égrõl" ("The Sun Is Up In The Sky"), dating back to the communist era. It predates all of the songs mentioned before and follows mainly the same melodic line during both variant phrases. Press the button on the left to listen to an excerpt; it is slowed down for the sake of comparison. It even has a shuffle beat that is comparable with the triplet-driven (or 3+3+2) lead melody of the chorus of "All About That Bass". What differs is a slower beat and rhythm. The triplets are just similar. The matching ratio is high: almost 100% of the school-song' notes appear in "All About That Bass" in the same timing.
  Now, did we find the original source? Probably not, because the school-camp song surely was not accessible enough outside of the Hungarian borders to inspire international Pop music. Moreover, if there's one such a song, there may be even earlier examples as well. As this case clearly shows there are basic melodic patterns that are driven by basic chord progressions, which may result in a long overlap, exceeding the eight note limit. Here, one surely cannot set claim on the melody because these basic melodies are as common as a cliché chord progression, while their origin (first use) is mostly unknown.
  There's a lesson to be learned from this: melody plagiarism is not always depending only on the length of the fragment but also on the singularity of its melodic shape and rhythm. To that, we can add, in many cases we will have to have additional factors to substantiate a claim of plagiarism. Take for instance the beginning of the chorus of Taylor Swift's "Blank Space" (2014) from the album 1989. This part of the song closely resembles the beginning of the verse of "Joyride" (Roxette; 1991). The fragment consists of seven notes — around the "unwritten limit" — packed in two bars of fast paced beats. The melodic design is simple and sparse (three different pitches).
  Of the songwriters — Swift, Max Martin and Shellback — at least one must have heard "Joyride", it being a well-known hit song. Unintentional inspiration? It may be so, but we can't be sure of it. In any case, as there are no additional factors amplifying this coincidence, we cannot prove plagiarism. Note that only the second chorus variant gives a perfect match. Moreover, there is a third variant as well that coincidentally offers a perfect match with a particular line from the song we discussed before: "Blurred Lines". And, this surely must be a coincidence (Table 6):
    1 . 2 . 3 . 4 . 1 . 2     : beats
        1 1 1   1 1 2   3     : "Joyride"
        1 1 1 1 1 1 2   3     : "Blank Space",
                                - first chorus
        1 1 1   1 1 2   3     : "Blank Space",
                                - second chorus
        1   1   1   2   3     : "Blank Space",
                                - "I leave you breathless"
        1   1   1   2   3     : "Blurred Lines",
                                - "I know you want it"
    1   .   2   .   3   .   4 : beats of "Blurred Lines"
  Table 6: "Blank Space" vs. "Joyride" vs. "Blurred Lines"
  For almost any song one can find some part that matches some part of another song. Take for instance the song, "Shake It Off" (2014) from the same album of Taylor Swift. It shows a rhythmical hook that sounds as if inspired by the sampler-hook of "Locked Out Of Heaven" (2012) by Bruno Mars (Table 7):
    1 . 2 . 3 . 4 . 1 . 2 . 3 . 4 . : beats
    * * *   *   * * *   *           : "Locked Out Of Heaven"
      * *   *     * *   *           : "Shake It Off"
  Table 7: "Shake It Off" vs. "Locked Out Of Heaven"
  And, "Locked Out Of Heaven", in turn, fosters a part that is seemingly borrowed from yet another song. The middle of the chorus introduces a melodic line (six consecutive notes) that perfectly matches with a melody in "Somebody That I Used To Know" (2011) by Gotye (feat. Kimbra) (Table 8):
    2 . 3 . 4 . 1 . 2 . 3 . 4 . : beats
    ... 1 3   1 6     5 5       : identical melody
  Table 8: "Locked Out Of Heaven" vs. "Somebody That I Used To Know"
  With its special syncopations and leap of a sixth interval this is a relatively characteristic melody.

Adding contestants to the competition. Delving in the history of popular music, for many matches between particular songs one can find other songs that match equally well or even better yet. The classic "Another One Bites The Dust" (1980) by Queen, for instance, has a signature bass hook. This hook shows similarities with the bass hooks of at least two other prior songs. The first one is "Good Times" (1979) by Chic. Like the Queen' song, this one starts with three E notes followed by a pause and then an "upbeat + downbeat" note on E. That sequence provides five identical notes in a row. The remainder is different, or just remotely similar, while the rhythm is very different. The simple "flat" shape of the consecutive common notes (all 1st degree) keep the sample out of plagiarism, but they are a very clear sign of inspiration, especially if we allow for the respective years of release.

  The bassline of the Queen' track is even more reminiscent of an early rap track called "Christmas Rappin'" by Kurtis Blow, also from 1979. This close similarity means identical rhythms and a 5/9 or 56% matching ratio. In spite of the three consecutive matchings only this one is closer to plagiarism, and an undeniable sign of inspiration (Table 9):
    1 . 2 . 3 . 4 . 1 . 2 . 3 . 4 : beats
    1   1   1      1112 3 4 5...  : "Good Times"
    1   1   1      11 1 3 14      : "Another One Bites The Dust"
    1   3   4      41 1 3 34      : "Christmas Rappin'"

    matching notes:
    *   *   *      **   *         : "Another One Bites The Dust"
                                    vs. "Good Times"
    *               * * *  *      : "Another One Bites The Dust"
                                    vs. "Christmas Rappin'"
  Table 9: "Another One Bites The Dust" vs. "Good Times" vs. "Christmas Rappin'"
  "Another One Bites The Dust" clearly is synthesizing the other two hooks with an 8/9 or 89% matching. Even the one note that is "missing" matches rhythmically. But separately handled it is not plagiarism. Curiously the second half of the "Another One Bites The Dust" hook (five notes) also appears in the bass hook of "Hey Bulldog" (Beatles, 1968). In spite of this coincidence and the fact that the songwriter (John Deacon) was admittedly a fan of the Beatles, the second half of the hook must have been inspired by "Christmas Rappin'". Now let's bring two more interesting samples into the comparison (Table 10):
    1   3   4      41 1 3 34  : "Christmas Rappin'"
    1   3   4     3   1       : "Smoke On The Water",
                                riff - second half
    1   5,  1      23  23  15 : "L'Arlésienne,
                                Suite No. 1 Overture"

    matching notes:
    *   *   *         *       : "Smoke On The Water"
                                vs. "Christmas Rappin'"
  Table 10: "Smoke On The Water" vs. "Christmas Rappin'" vs. "L'Arlésienne"
  The first sample again is from "Christmas Rappin'". The bassline of this song shows a 4/5 or 80% matching with the second half of the legendary riff of "Smoke On The Water" (1972) by Deep Purple. Note that the same ratio is only 4/9 or 44% as calculated from the nine notes of the "Christmas Rappin'" sample. Except for the first three notes the samples do not sound close to each other in the second bar. For the second sample we travel a bit farther back in time. As it happens "Another One Bites The Dust", "Christmas Rappin'" and "Good Times" all show a 7/9 or 78% rhythmical matching with the overture "hook" motif of "L'Arlésienne", Suite No. 1, by Bizet, dating back to the early 1870's. Bizet's tune, moreover, shows a 4/9 or 44% melodic matching with "Good Times".
  "Get Lucky" (2013) by Daft Punk has a special chord progression of unknown origin: iv —> VI —> i —> VII. It is a turn-around progression, but probably one of the latest among the members of this genus of progressions. At least three recently released songs using the same progression have been competing for the prize of being the original. Each one's case was supported by a dedicated mash-up video on the internet.
  The first of these songs is "Get Out" (2011) by Milks. Their mash-up video was a cross-fade mix of the two songs, looping only the instrumental hook. The crossfade transition was very smooth, trying to prove that both songs are substantially the same. While the mash-up track is very convincing, we have to be careful with mash-up mixes because crossfade-trick will always work fine with songs built upon identical chord progressions and similar arrangements. "Get Lucky" could probably as easily be mashed up with the "Night Fever" (Bee Gees) groove as well. Some people even submitted a mash-up mix of "Get Lucky" (backing) and Michael Jackson's "Billie Jean" (vocals).
  The second song was an instrumental track from the Korean guitarist Zack Kim by the title of "Robot Dance" (2011). This mash-up mix that was submitted to YouTube also worked fine, since the chord progression and tempo (bmp) were adjusted.
  The third song is "Love Life" (2013) by John Mamann. It has a lead melody that is very similar to the pre-chorus of "Get Lucky". At a first listen both songs really do sound almost the same. A closer look, however, shows the differences as well. Both "Get Lucky" and "Love Life" are using a resembling, though slightly different rhythmical pattern (Table 11):
    1 . 2 . 3 . 4 . 1 : beats
    *         *  *  * : "Love Life"
    *         * * *   : "Get Lucky"
  Table 11: "Get Lucky" vs. "Love Life": rhythmical pattern
  This slight difference considerably decreases the ratio of perfect matchings between the samples. One has to admit, however, that such a close similarity has to be counted as adding up to plagiarism. The melodic shape itself is not perfectly the same either in the second and the fourth sub-phrases. The closing notes of the fragments are paralleling the roots of the actual chords in thirds (Table 12):
    6, > 6,7,1 > 1 2 3 > 5 4 3 2 : "Love Life"
    6, > 5,6,1 > 1 2 3 > 3 4 2   : "Get Lucky"
  Table 12: "Get Lucky" vs. "Love Life": melodic shape
  Luckily for "Get Lucky", the identical timing matching ratio is only 3/11 or 27% for the melody, and comes out at 7/11 or 64% for the rough melodic shape. On the other hand, the similarity is stronger than what this ratio suggests. It is massively enhanced by the same chord progression being used (be it in different keys) and some other close matchings. Now let's bring yet another sample into the comparison: the chorus of "Say, Say, Say" (1983), a duet of Paul McCartney and Michael Jackson (Table 13):
    6, > 6,7,1 > 12321 > 321...  : "Say, Say, Say"
  Table 13: "Say, Say, Say": melodic shape
  The details this song shares with "Love Life" are the melodic line that's rising up to the 3rd degree and the rough timing of the sub-phrases. This similarity is as close as the one with "Get Lucky". The main difference lies in the chord progression where "Say, Say, Say" omits the closing VII chord and at the end of the melody line where the closing note is also different.
  A similar long-running "rip-off contest" started with Coldplay's hit song "Viva La Vida" (2008). People on the internet noticed that this song's verse in many respects equals the chorus of Joe Satriani's track "If I Could Fly" (2004), both in its melody and in its chord progression. Next, some of them found out that "If I Could Fly" is reminiscent of a whole series of other melodies, ranging from an earlier Cat Stevens' song-suite called "Foreigner Suite" (1973), over "Francés Limón" (2002) by the Argentinian band Los Enanitos Verdes to "Hearts" (1981) by Marty Balin. The verse of the Beatles' song "You Won't See Me" (1965), by the way, also starts similarly, though with quite different chords.
  The chord progressions in "Viva La Vida" and "If I Could Fly" indeed closely resemble one another: IV —> V —> I —> vi, and: ii —> V —> I —> vi respectively. The pair of chords that differ are relatives — chords that are easily interchangeable in the idiom of Pop and Rock music. A note-by-note comparison shows that the first bar with those characteristic three notes motif is the same indeed, but the remainder of the melodies is rather different. The closely resembling chord progression, and two anchoring notes of the melody keep the melody sounding like a close variant, even though it's not that close at all. The matching ratio comes out at 5/14 or 36%.
  The "Francés Limón" sample — like "If I Could Fly" a solo guitar tune — perfectly matches the first bar, while two further anchoring notes are also matching (Table 14).
       Bar 1-2
       1 . 2 . 3 . 4 . 1 . 2 . 3 . 4 . : beats
    ...3           4 2             2 x : "Viva La Vida"
       3           4 2               1 : "If I Could Fly"
       3   4   2   1 2       7,7,1 2 2 : "Foreigner Suite"
       3           4 2         7,1 2   : "Francés Limón"
       3           4   2...            : "Hearts"
    ...3           5 2...              : "You Won't See Me"

    matching notes:
       *           * *                 : "Viva La Vida" vs.
                                         "If I Could Fly"
       *           * *                 : "If I Could Fly" vs.
                                         "Francés Limón"
       *             *                 : "Foreigner Suite" vs.
                                         "If I Could Fly"

       Bar 3-4
       1 . 2 . 3 . 4 . 1 . 2 . 3 . 4   : beats
       2   2 1 4   6,7,  3 3 3 3 3 3 3 : "Viva La Vida"
       2     1 2   2 1 2 3   1   6     : "If I Could Fly"
         3   2   1 2 6,                : "Foreigner Suite"
       2           3 1                 : "Francés Limón"

    matching notes:
       *                 *             : "Viva La Vida" vs.
                                         "If I Could Fly"
       *             *                 : "If I Could Fly" vs.
                                         "Francés Limón"
                   *                   : "Foreigner Suite" vs.
                                         "If I Could Fly"
  Table 14: "Viva La Vida" vs. "If I Could Fly" vs. "Foreigner Suite" vs. "Francés Limón" vs. "Hearts" vs. "You Won't See Me" (transcribed samples are taken from the excerpts submitted on YouTube)
  The ratio is 5/9 or 56%, the closest fit among the comparisons. "Foreigner Suite" is no real contestant in this comparison due to a different melodic shape and a deviant rhythmical pattern. Still it's also remotely reminiscent due to a similar chord progression and just a 2/3 matching of anchoring notes.

Guilty of the offence? Now let's take a look at some songs that — rightly or wrongly — were found guilty of the offence of plagiarism. The legendary guitar intro of Led Zeppelin's "Stairway To Heaven" (1971) was judged to have plagiarized the Spirit's instrumental called "Taurus" (1968). The charge regards the three initial bars — not counting the long keyboard intro of "Taurus". Listening to the "Taurus" sample, one really can notice the similarity with "Stairway ..." Many people can even name the latter song without any hint. The opening three notes are exactly the same; played on the same instrument (acoustic guitar), in the same tempo and cultivating the same sound. On the other hand, the use of an 1-3-5 arpeggio pattern in a minor key is a platitude, similar to the choice of the a-minor "homekey" in respect to guitar ballads.

  Next to this, the "Taurus" arpeggio pattern remains rising, while "Stairway ..." changes into a falling pattern and adds a counterpart on top of that. This shape narrows down the resulting matching notes to one bass note per bar. The arpeggios are outlining the minor tonic chord combined with a chromatic descending line. This concept is a cliché, see for example "Chim Chim Cher-ee" from the film Mary Poppins (1964). This latter example comes even closer to "Stairway ..." because the fourth bass note is building up to the subdominant chord (D), while "Taurus" sustains the initial concept for the fourth bass note as well. Another point of similarity is the stopping descent of the bass in the fifth bass cycle.
  Out of the 30 notes of the intro bars of "Taurus" 9 are identical with those of "Stairway ..." or 30%. Five of those, moreover, are coming from the same bassline cliché being followed. A further four weak matches can be added "compensating" the different arpeggio pattern — weak, since the matching of arpeggio notes is an easily earned point. Note that the matching ratio with a simple 1 —> 3 —> 5 —> 8 arpeggio pattern looped would be 11/30 or 37%, no less.
  If we do the same test for "Man Of Mystery" (1960) by The Shadows vs. "Taurus", the matching ratio comes out at 7/30, or 23%. See the guitar tablatures of the three samples to identify the matching points (Table 15):

    "Stairway To Heaven"

    "Man Of Mystery" (transposed)
     1 . 2 . 3 . 4 . 1 . 2 . 3 . 4 . 1 . 2 . 3 . 4 .

    "Cry Me A River"
     1 . . 2 . . 1 . . 2 . . 1 . . 2 . . 1 . . 2
  Table 15: "Stairway To Heaven" vs. "Taurus" vs. "Man Of Mystery" vs. "Cry Me A River"
  And, here are the matching notes (Table 16):
    matching notes:
     * *     *       *       *   *   *           : "Taurus"
                                                   vs. "Man Of ..."
     * * *   *       *       *       * *   *     : "Stairway ..."
                                                   vs. "Taurus"
     * *     *       *       *       *           : "Stairway ..."
                                                   vs. "Man Of ..."
     * * *     * *     * *     * *     *       * : "Taurus" vs.
                                                   1 > 3 > 5 > 8
                                                   pattern looped
  Table 16: "Stairway To Heaven" vs. "Taurus" vs. "Man Of Mystery": matching notes
  In 1959 the British guitarist Davy Graham performed an instrumental cover of "Cry Me A River", written by Alex Hamilton in 1953, for a BBC television documentary. Graham's version, also released on his album The Guitar Player (1963), is another prior example that uses the same cliché progression, though in 6/8 meter. It also features the characteristic 7-> 8 -> 7 melodic motif played on the top string in the slow-down measure combined with a slide up which is familiar from the live versions of "Stairway To Heaven". Definitely the accessiblity is more evident in the case of the Spirit' song, as both bands had been touring together back in 1968. This fact increases the probability of inspiration. On the other hand, there is not much detail in "Taurus" that is absent from "Cry Me A River" but present in "Stairway To Heaven": only the tempo and the 4/4 meter. The arpeggiated chords of "Taurus" come much closer to "Cry Me A River" than to "Stairway To Heaven", especially if we include the repeated 9th note on top of "Cry Me A River" that clearly resonates with the 2nd degree built in the "Taurus" arpeggios. Considering the similarities between "Cry Me A River" on the one hand and "Taurus" and "Stairway To Heaven" on the other, it does not seem to be reasonable to share the songwriting credits of "Stairway To Heaven" with "Taurus" songwriter Randy California, as the court judged in 2014.
  All in all, "Stairway ..." may be called a clear case of inspiration but not of plagiarism. A better example of the latter is George Harrison's song "My Sweet Lord" (1970). The song sounds very reminiscent of "He's So Fine" (1962) by the Chiffons. Even at a first listen, one can easily hear the similarities. Let's see the melodic matchings (Tables 17-20):
    1 2 3 4 1 2 3 : beats
      4 4 4 5 55  : backing vocal hook #1 "He's So Fine"
      4  6 65  5  : backing vocal hook #1 "My Sweet Lord"
      *     *  *  : matching notes
  Table 17: "My Sweet Lord" vs. "He's So Fine": backing vocal hook #1
    1 2 3 4 1 2 3 : beats
      5 5 5 5 55  : backing vocal hook #2 "He's So Fine"
      5  5 6   6  : backing vocal hook #2 "My Sweet Lord"
      *           : matching notes
      5 55 6      : "Oh, Happy Day"
  Table 18: "My Sweet Lord" vs. "He's So Fine": backing vocal hook #2
    1 2 3 4 1 2 3 : beats
      5 5332      : verse motif #1 "He's So Fine"
      5 3 32      : Verse motif #1 "My Sweet Lord": "oh my Lord"
      *   **      : matching notes
  Table 19: "My Sweet Lord" vs. "He's So Fine": verse motif #1
    1 2 3 4 1 2 3 : beats
     55616161 1   : verse motif #2 "He's So Fine"
       556161 1   : verse motif #2 "My Sweet Lord": "I really
                    wanna ..."
         **** *   : matching notes
       **         : closely matching notes
  Table 20: "My Sweet Lord" vs. "He's So Fine": verse motif #2
  Now, looking at the matching ratios, the case for plagiarism does not seem as convincing as before. The comparisons show distantly similar and closely similar motifs as well with up to five consecutive notes matching in a two noted not syncopated pattern. Still, these four fragments are reinforcing each other effectively. The reinforcing effect works strongly because the samples are neighbouring each other, sometimes even overlapping, and their structure is not generic. The mash-up comparison between both songs works effectively for a longer section of the song.
  The first notes of the lead vocal motif (verse motif #2) do not match, but one feels that they are coming close to each other. Does the "You Really Got Me"-effect that we described above, not work here? It seems that the shift by one quarter in time changes the melody's character less than the shift by one eighth, as the former keeps the accent on or off-beat. The second backing vocal hooh (hook#2) with a matching ratio of only 1/6 (17%) also sounds reminiscent, due to the fact that the backing chords are changing in the same way and the rhythm sounds like a syncopated variant of the motif in "He's So Fine".
  The song "Oh Happy Day" by The Edwin Hawkins Singers (1969), by the way, is another possible candidate here to be named as the source of inspiration. Some chord changes and the "backing vocal hook", and the antiphonal arrangement of these backing vocals do in fact sound rather similar, but considering the other similarities one can say that "He's So Fine" must be the real source of "My Sweet Lord".

Adding modifications. The ostinato hook of the Nirvana song "Come As You Are" (1992) sounds rather similar to the intro of "Eighties", a song by the group Killing Joke and dating back to 1983. This is yet another similarity that is instantly recognizable. Not just the notes but also the functions are close to each other: the intro as well as the verse hook. The special guitar tone, moreover, makes the songs sound alike (Table 21).

The matching ratio, however, is only 6/14 or 43%, also counting the "spike" note (the 4th degree) that jumps up one octave in "Eighties" by the side of "Come As You Are". This score is a surprisingly low value given one's first impression. From the second beat on there are four oscillating notes with an inverted pattern. On paper this offers a zero matching, but our ear suggests that the inverted oscillation is also a case of similarity and can be considered to be a close matching. Some analysts may derive this subtle similarity from the samples being retrogrades (reversed): 3-1-3-1 vs. 1-3-1-3. Similarly, the notes falling on the last eight of the first bar are also coming close-by — 7 vs. #7, a passing tone. In the middle of the second bar we have a 4' —> 7 sequence shifted by one eight (marked with //), which adds yet another weak matching. Together with these "imperfect" matchings we have already 13 consecutive notes, all combined with the special guitar tone. Among the matchings we have two (+ two) leaps over the interval of a third.

    1 . 2 . 3 . 4 . 1 . 2 . 3 . 4 . 1  : beats
    1   3 1 3 1 1#7,7,4 7,7,  4 7#7... : "Come As You Are"
    1   1 3 1 3 1 7,7,4'7,7,4'7,3'3    : "Eighties"
    *           *   * - * *            : perfectly matching notes
        * * * *   *   *     / /        : closely matching notes
    3   5 3 5 3 3 3 2 6 2 2   6 2 3    : test melody
  Table 21: "Come As You Are" vs. "Eighties"
  Now, take the "test melody" in Table 21 that has no matching notes at all in common with "Come As You Are", but is roughly paralleling it. With the same guitar tone it would also instantly evoke a feeling of likeness. Here the question arises if we can create a melody that is instantly and strongly reminiscent of another piece of music, but that is not plagiarizing it? We will keep this this question open. Or, what about making a major key to minor key conversion or a straight beat to shuffle beat conversion copy?
  And, what happens if we copy a particular melody with say 15 consecutive notes and underlying chords, then start to modify it by altering the pitches one by one, and then gradually change the rhythm and the chords too. Soon we will obtain a melody that is already falling into the safe area of non-plagiarism. It still may sound slightly reminiscent to the original, especially if the accompaniment is also kept similar. According to the "sensible inspiration = plagiarism" approach you are not safe from the charge of plagiarism until any reminiscence of the source has been removed. After further minor modifications the substantial similarities fade away and we reach a point where an outsider cannot notice anymore that the two melodies are related. A few more modifications will do away with all of the subtle similarities.
  Keep in mind, moreover, that a few subtle — in legal terms: de minimis — similarities always can be found between independently created melodies. Think of the apples and oranges that are famous for being too different to be compared. In spite of this, we can find substantial similarities between the two: both are round-shaped fruits, lay close-by in the colour range (red ≈ orange), reproduce by internal seeds, and so on ... We saw for example that relatively short 4-note consecutive not quite identical fragments still can be strongly reminiscent to another, if the similarities are reinforced effectively by a sufficient amount of imperfect matchings and a special distinct tone. So, one will have to be careful with imperfect matchings, because depending on the ease with which we allocate them, we can match very subtly and slightly resembling motifs as well, and turn almost any pairing of melodies into a "substantially similar" one.

Conclusions. What has to count in cases of plagiarism in Pop and Rock music? Subtle similarities, substantial similarities, identical melodies and details, or rather specific combinations of these? Unfortunately the law does not answer this question directly, so the opposing parties at a trial may easily re-define these classes to their own ends. Subtle similarities can be argued to be substantial, similarities can be enhanced to approach identity and vice versa. Even a simple comparative test can be manipulated to promote superficial similarities into substantial or identical ones. Lay listeners, jurors and judges alike, can easily be misguided by all the musicological talk and trickery involved. It's therefore the responsibility of the musicologist experts of the opposing parties to debunk cheats and to force out a common and intelligible conclusion on most of the major points before the jurors are asked who's right and who's wrong. For the right decision prior cases and well-chosen examples may be of help, especially on questions of what is generic and what is unique. The existence of predating examples will usually further the defendants' case. Unfortunately for the plaintiff, a lack of examples to be found will usually not prove the non-existence of such example.

  Can the work of experts made easier and can the comparative method be turned into an algorithm that includes variables like imperfect matchings, unusual or usual melodic shapes, rhythms and chord progressions? Can the class of "substantial similarities" be narrowed down? Unless these questions are clarified, judgements will inevitably tend to be partisan. There are already working algorithms for melody recognition that are combined with an extensive database of music (e.g. midi-files). There have been experiments running for implementing a software detecting signs of plagiarism. However, since the criteria are hard to compile and even hard to get accepted, quick results are not to be expected.
  Finally, think of the fact that coming up with a fully original melody nowadays is far more difficult than it was about fifty or sixty years ago. And, it's getting "worse" by the decade. Back then the introduction of expanded modal melodies and chord material and new types of syncopation enabled songwriters to create more original melodies over new types of chord progressions, rhythmical patterns, instruments, tones and production techniques. Since then the melodic, harmonic and rhythmical arsenal has not seriously boomed up. The evolution of musical genres has also slowed down, especially in the past two decades. The volume of the musical production, on the other hand, has multiplied. These opposing movements have seriously increased the probability of accidental matchings in every aspect of songwriting. Sooner or later these phenomena will have to be accounted for in the approach to plagiarism by admitting more matchings of generic patterns.
1. See: US Copyright Law, Copyright Act, 1976. Return to text
2. C —> Am —> F —> G in the key of C-major. Return to text
3. See for an example the song "Man Of Mystery" (1960) by The Shadows. Return to text
4. D —> E —> G —> D chords in the key of D-Major. Return to text
5. There's no perfect mathematical model describing the variation of melodies. Return to text
6. This method was developed by Billings Learned Hand. Return to text
7. Read the great and detailed analysis of Ian "paramucho" Hammond (1999). Return to text
8. See: Finell (2013). Return to text
  Relevant websites
  2015 © Soundscapes