For me, the alignment of my characters is a way to help me think about how to roleplay them, but it's just part of a mosaic and I don't view it as a restriction. I don't think in terms of "my character wouldn't do this because he's lawful good" or "my character is doing this because he's chaotic neutral." It's just a way to start thinking about personality, but not a straight-jacket. As a DM, I would never try to correct or corral a PC's behavior because it doesn't "fit" their alignment, or tell them, "If you keeping doing stuff like this, your alignment will change to x."
I don't expect D&D to be a simulation, so my complaint about viewing alignment rigidly isn't that it's not realistic (LOTS of stuff in D&D isn't realistic and doesn't bother me). What bothers me about alignment being rigid or polar is that it's inherently undramatic. It leads to less psychologically nuanced/complex characters. I don't think it's good storytelling to have a person who views every situation they're in through SUCH a specific lens. A person might behave honorably in one situation and selfishly in another due to a complex variety of factors.
That said, one DM I know was able to give a me great example of a group dynamic where emphasizing player character alignment was very helpful in keeping the game fun for the table. The player was having their character do something that was grossing the DM out (harvesting body part trophies) so the DM started a conversation with him about it by asking the player if their character was "evil." It lead to the revelation that the player was taking the body parts to set up a weird elaborate joke down the line, and wasn't aware that it was bothering people, and agreed to just stop doing it because it hadn't occurred to him that it was "evil." So for some groups, it can obviously be a great tool.
I think the old convention of alignment can be used in combination with the newer ideas of bonds, traits, and flaws to help you form an idea of what your character is like when your first start out playing them. For me, after that point, I find that characters I play tend to tell me who they are - much like when I'm writing a script or a story - and can sometimes surprise me by turning out quite differently that the way I "planned" them to be.
Having said all that, I was amused in running Curse of Strahd to encounter a couple of moments when alignment DID mechanically matter in a very old-school way. In both instances, the importance of alignment was left in the adventure as an artifact of the original 1E module - there are several places in Curse of Strahd where text has just been copied verbatim from the original Ravenloft adventure (including some places where the 5e team later regretted it, such as the description of the Vistani). Given a less frenzied production process, I wonder if those instances of alignment being very mechanically important in a very non-5E way would have ended up excised. I kind of liked them being in there because it just felt VERY different, but I wouldn't want to be dealing with cases like that often.
I think Eberron (first published in 2004) was the first official D&D setting to heavily downplay alignment, and especially the idea of attaching default alignments based on a creature's race. Pretty much every intelligent creature in Eberron is capable of great good or great evil on an individual basis. The only creatures that adhere to alignment based on the type of being they are are basically outsiders like celestials or fiends who effectively embody various moral concepts. I think 5E in general has taken its cues from the Eberron model, and I expect things to go even further in that direction in the next few years.
I don't expect D&D to be a simulation, so my complaint about viewing alignment rigidly isn't that it's not realistic (LOTS of stuff in D&D isn't realistic and doesn't bother me). What bothers me about alignment being rigid or polar is that it's inherently undramatic. It leads to less psychologically nuanced/complex characters. I don't think it's good storytelling to have a person who views every situation they're in through SUCH a specific lens. A person might behave honorably in one situation and selfishly in another due to a complex variety of factors.
That said, one DM I know was able to give a me great example of a group dynamic where emphasizing player character alignment was very helpful in keeping the game fun for the table. The player was having their character do something that was grossing the DM out (harvesting body part trophies) so the DM started a conversation with him about it by asking the player if their character was "evil." It lead to the revelation that the player was taking the body parts to set up a weird elaborate joke down the line, and wasn't aware that it was bothering people, and agreed to just stop doing it because it hadn't occurred to him that it was "evil." So for some groups, it can obviously be a great tool.
I think the old convention of alignment can be used in combination with the newer ideas of bonds, traits, and flaws to help you form an idea of what your character is like when your first start out playing them. For me, after that point, I find that characters I play tend to tell me who they are - much like when I'm writing a script or a story - and can sometimes surprise me by turning out quite differently that the way I "planned" them to be.
Having said all that, I was amused in running Curse of Strahd to encounter a couple of moments when alignment DID mechanically matter in a very old-school way. In both instances, the importance of alignment was left in the adventure as an artifact of the original 1E module - there are several places in Curse of Strahd where text has just been copied verbatim from the original Ravenloft adventure (including some places where the 5e team later regretted it, such as the description of the Vistani). Given a less frenzied production process, I wonder if those instances of alignment being very mechanically important in a very non-5E way would have ended up excised. I kind of liked them being in there because it just felt VERY different, but I wouldn't want to be dealing with cases like that often.
I think Eberron (first published in 2004) was the first official D&D setting to heavily downplay alignment, and especially the idea of attaching default alignments based on a creature's race. Pretty much every intelligent creature in Eberron is capable of great good or great evil on an individual basis. The only creatures that adhere to alignment based on the type of being they are are basically outsiders like celestials or fiends who effectively embody various moral concepts. I think 5E in general has taken its cues from the Eberron model, and I expect things to go even further in that direction in the next few years.