I noted in a thread on the solr-user mailing list https://lists.apache.org/thread.html/557347c2e22352ae3b2ccd9e9f22fbf8ce50d8efed901901adf020fc@%3Csolr-user.lucene.apache.org%3E that token graph filters (e.g. SynonymGraphFilter and WordDelimiterGraphFilter) can't handle input graphs, i.e. streams produced by other token graph filters.
From the above-linked thread, Webster Homer wrote:
I noticed that in some of the current example schemas that are shipped with
Solr, there is a fieldtype, text_en_splitting, that feeds the output
of SynonymGraphFilterFactory into WordDelimiterGraphFilterFactory. So if
this isn't supported, the example should probably be updated or removed.
We should evaluate all analysis chains in shipped schemas to address this.