This post is also available in: German
To better understand the 又 example, here is some more background information.
Diagram C2.1 shows the evolution of 又 yòu “right hand”. In (a) the arm is on an up-down axis and the fingers fan out towards the left. No, the ancient Chinese did not only have three fingers. This is basically a stick-figure representation of a hand. In (b) the fingers all point up and and to the left. (c) is a much more stylized version and the arm is now leaning at a 45 degree angle. In (d), the top finger is bending towards the middle finger, which end up connecting in (e), making (e) look much less like a hand.
Diagram C2.2 shows the evolution of 左 zuǒ “left hand”. Here (a) is a left hand. Note that its orientation is opposite to that of (a) in Diagram C2.1. Most of the time, changing a component’s orientation didn’t change its meaning in early scripts like oracle bone and bronze inscription scripts, but this is a case in which orientation does make a difference. In (b), 工 gōng “work” is added to the form (For now, I’m going to ignore the “why” 工 was added as it would take several more paragraphs to explain and its explanation isn’t pertinent to this post). In (c) through (e) you can see how the form becomes more and more stylized.
Diagram C2.3 shows the evolution of 友 yǒu “friends”. Though this character was originally a picture of two right hands, it is easily seen in (d) and (e) that each “right hand” evolved separately, which is in and of itself an important thing to note: 1. Even though two forms start out the same, they may diverge over time. and 2. The opposite is also true: things which started out different may converge over time. If you compare the evolution of 左 in Diagram C2.2 and 友 above, you will notice that the “right hand” on the right in (a) & (b) and on the top in (c), ends up looking exactly like the “left hand” from Diagram C2.2 (e). Or, you could just compare 左 and 友! The fact that calligraphers generally write the top parts of these two characters differently is directly related to their separate origins.
Putting points #1 and #2 a different way: Character components which used to look the same, may look different now and character components which look the same now may have been very different at an earlier time. Basically what that means is, we can’t look at a modern form and know anything about it with certainty without first doing the research. This is exactly why we want to make a dictionary (beyond the fact that it’s incredibly fun and interesting), so that you don’t have to do the research!
Back to the post:
Individual characters usually have several variants. These variants often co-existed, each having the possibility of being loaned to another word because of its sound, having further components added to it, subtracted from, having its components melt together, dying off or just going on its merry little way. After characters became more standardized (albeit rather loosely), there were fewer variant forms, but variant forms will always be around.
So, the relationship between and is that the latter is a variant form of the former. It’s a variant form that had already died out by the time of the Shuōwén. These forms probably co-existed for some amount of time. So, the form would have most likely evolved into if it had not been corrupted into, because the top part of is which did evolve into 友 as seen in Diagram C2.3 above. The bottom part of is 甘, hence the form.