[Deduplication] What about resources that fit into more than one category?

Came across this issue while designing the information portal deduplication UI part: Should the duplicate suggestion algorithm only search within one category or across all categories? What happens when a resource exists multiple times in different categories (e.g. a repository as a repository and as a data source, or a vocabulary as a terminology and as a metadata-standard)?

My currently favorite idea for dealing with this would be to have such resources multiple times, one for each category and links between them via an optional field:

"is also a": [
    "<link to same resource in other category>",
    "<link to same resource in yet another category>",
    ...
]

Another option would be to allow single resources to belong to multiple categories at the same time. This may or may not be more complicated with regards to validation.