(Here's the '54 Bowman Marshall card for illustration)
I'd like to share that here today:
My standard answer is that if the card was never corrected, it really has no extra value as a result. For instance, look at the 1959 Lew Burdette card. It shows Lew as a lefty (he and Warren Spahn swapped gloves on picture day as a practical joke), but the picture isn't a true error. Even then, Lew's name is spelled "Lou" on the card (which Topps repeated in '60). Neither the photo nor the spelling of his first name were changed, so the card has no added value as a result.
On the other hand, there's the 1957 Topps card of Hank Aaron with a reversed negative. A quick look at Aaron's uniform number shows it's backwards. This card has added value over common 1957 Topps cards, but not because it's an error. I'll argue that the 1957 Topps card of Aaron wouldn't be any less valuable today if the negative had been corrected and reversed before the cards were printed...of course, correcting the photo afterwards would have created a different variation and the hobby would have determined whether one variation would be more valuable than the other: like the 1958 Yellow variations and the 1969 Mickey Mantle variatons.
I guess the short answer to your question would be that if the card never was corrected, there really isn't a reason to worry about it as far as the hobby is concerned. Otherwise, we'd be having to list every single "hatless" card of the 1960s, every card that wasn't airbrushed before it was printed (like Don Zimmer's 1962 Topps card with him in a Mets uniform but listed as a member of the Reds) and every card with a typo on the back as an "error" card.
If the collector in question can produce a legitimate 1954 Bowman card of Willard Marshall in a different uniform...then he'd have something.