"influenced by" artist-artist rel (STYLE-1615)

Tags: #<Tag:0x00007f0eb1353e48>

On STYLE-1615 @Zas proposes an “influenced by” relationship that would allow linking artists to other artists they claim as their influences, or influences that are “well-known (or obvious)”.

I find this information interesting, but I’m afraid about the trustworthiness of the data. I’d expect us to only allow direct claims from the artist (or their press texts) if we were to add something like this, and it would need to be understood as a claimed influence - I can claim I’m influenced by Iron Maiden when making music with my vacuum cleaner without anyone else hearing anything similar to Iron Maiden in there, after all.

What do people think about such an idea?

4 Likes

I think it’s highly subjective, even when an artist explicitly states the influence. On the other hand we already store data about genres, which is equally subjective.

In other words I’m not sure about this relationship.

3 Likes

Main difference between genres and a relationship is that the genres are tags and separate from the relationship system - relationships should be considered objective enough that they can hopefully consistently be voted on one way or the other :slight_smile:

Another thing to consider is that artists may be reticent to name their influences for legal reasons. Ever since the infamous “Blurred Lines” copyright case, the music world has been on edge, worried that any named influence or inspiration could be used against them in a plagiarism lawsuit. To this end, it may not be a good idea to have any such ARs.

1 Like

If an artist X publicly says he was inspired by band Y, or if a writer, specialist of related genre or band X says so, and if editors agree, then we can just create a relationships.
Copyright cases have and will happen with or without such relationships in MB.
To me, that’s totally unrelated.

Of course it is, but they are plenty of subjective informations around that are still useful, right?

For example, Metal Archives has lyrical themes for bands, I guess that’s very subjective, but that’s clearly a nice addition to genres (in the same genre, you can have various lyrical themes), as it helps to define the band.

Wikipedia is also listing lyrical themes in their articles (example: https://en.wikipedia.org/wiki/Pink_Floyd#Lyrical_themes).

It’s more about the process, and we clearly do that already with genres.

As other data, any relationships should be sourced, and voted.
Example for sources related to influences: https://en.wikipedia.org/wiki/Pink_Floyd#cite_note-425

Another partially subjective thing Wikipedia is listing is Associated Acts (https://en.wikipedia.org/wiki/Template:Infobox_musical_artist#associated_acts)

We don’t have those (as is), but they exist on Wikipedia because users/editors found this kind of information useful.

AFAIK, MusicBrainz database is made by people for people, and as a music fan, an editor, and a human I think knowing that artist X was influenced by artist Y is a useful information, hence the ticket.

One may disagree about the usefulness, but no one is forced to use or edit this data anyway.
Same goes for genres.

Also to be honest, we have plenty of perfectly objective data in the database that is totally useless to most users.
As an example, I don’t care if artist X is married to artist Y, but perhaps if those artists are musically influencing each other, or writing music together, which may not be the case at all.

Though this relationships exists in the database, perhaps because some people found it useful.

I don’t ask to remove such “married to” relationships, I’m perfectly fine with it, I even added some of those relationships.
But if you asked me to choose between “married to” or “influenced by” relationships, in the context of a database dedicated to music, I would have chosen the latter right away, just because the former is unrelated to music.

5 Likes

I do like such new ideas, and while I probably would not use this, if other users find this useful, why not.

There are two possible issues that come to mind.

First ‘what inspired the artist’ is quite vague.
Who inspired him at seven years old to pick up a guitar?
Who influenced the music on his first album when he was 20 years old? Who inspired the album he released when he was 50 years old?
The influences could even be different for different songs on one release.

Secondly, I would not like these relations showing on the main page of a release.
Even the ‘sampled by’ information already annoys me a bit :wink:
(and those are even facts)
Perhaps a link on the side could be fine? Not sure.

1 Like

If we only added such relationships in the cases where there is verifiable information (i.e. “artist intent”) it wouldn’t be problem, but I definitely think a relationship like this will lead to editors doing highly subjective and arbitrary classification on artists even in the absence of such information. There is a reason we do not have genres as relationships, but instead as a very clearly subjective and moderately democratic tagging system. This type of subjective data could also easily lead to edit warring, which is not a fun activity for anyone involved and quickly leads to burning out well-intentioned contributors.

And given the above I also question the usefulness of such a relationship, it will either be too rarely entered and incomplete for data consumers if we only follow official artist intent, or too arbitrary if we allow editors to enter subjective data.

3 Likes

More information that people can add is always a plus in my opinion. I personally wouldn’t really use it but if someone would like to then great.

I think the guidelines the reosarevok has already laid out addresses any issues about the data being subjective. If the artist saying something is ‘subjective’ then so is when they say when they recorded a song etc, but we would of course store that information. Also, ‘influenced by’ is clearly a more vague piece of data than some others - nobody is being misled and I don’t think any database that is linked into MB will collapse if there’s disagreement on what to put into that relationship for an artist.

edit: and data like this is that is reasonably tucked away for now could eventually become something more, like a ‘also check out this artist’ panel on MB or LB, and so on

Imagine the shock when Steve Vai was inspired in his early youth by Django Reinhardt, and somebody gets both in his playlist.
This actually could be fun :wink:

1 Like

The problem is how to enforce such guidelines, we already have problems with that. For example: Does Wikipedia count as an official source? If yes, that’s already a basically infinite source of disagreements.

3 Likes

Quite frankly, if it’s widely accepted to be a non-concrete (yet useful) datapoint, and it doesn’t break anything, who cares?

This all just feels like a bit of “I wouldn’t use it but I can imagine it not being perfect so no”.

1 Like

But can anybody seriously challenge the fact Steve Vai was influenced by Frank Zappa and Joe Satriani?
Wouldn’t be a valuable information to have for people interested in this artist?

This is a good example of why this could become very interesting, we’re already fighting :wink:

While the idea surely is interesting and has validity, to me there seems to be some blurred lines between ‘influenced by’ and ‘sounds similar to’.
Is this aiming to: “If you like this, you might like that” (= sounds like)
Or do we want to archive that Steve Vai once mentioned in an interview that Django Reinhardt was the reason he dumped his recorder (small basic flute) in the trash can and stole a guitar?

Neither.
You are deviating here. Vai says he’s influenced by Zappa (and he did more than once over 40+ last years), we just store that, simple.

Ok.
So you mean that the only valid reason for entering an ‘influenced by’ entry is if there is proof that an artist has stated somewhere that he has been influenced by somebody?

That would mean that if an artist has played in a band when he was younger, but never mentioned it as an influence, it should not be entered as ‘influenced by’?

And if an artist is creating music that (obvious to the listener) was influenced by another artist/composer, but never mentioned that anywhere—and perhaps was not even consciously aware of it—also doesn’t qualify?

Read again the ticket, it starts with:
“When an artist specifically claims another as an influence, or when this influence is well-known (or obvious).”

It seems to me it’s clear enough. You can always discuss what’s “well-known” or “obvious”, but basically it means with sources and/or references (those can also be discussed etc…).

1 Like

My concern is that I do not want to have to go into edits on the artists I subscribe to to potentially start conflicts debating such issues.

3 Likes

Yes, so that confirms that you mean entries would only be valid if there is proof that the artist himself has stated somewhere that he has been influenced by artist x.

And any obvious influences (to the listener) that can not be proved by the artist acknowledging this himself doesn’t qualify.

In that case I am not very interested in what this proposal will result in.

I would also like to petition getting rid of a lot of relationships that I don’t use and cause debate (the whole classical guidelines really), because I get a lot of notifications, and I’m not interested.

Filtering notifications via blocking additions to the DB (rather then actually using the filters) that others would find useful is not very team spirited.

1 Like