tatum() behaves oddly in presence of **harm spine

[musicnerd]

running tatum() on the bhchorales gives extremely odd output, likely because it is assuming recip info in the **harm spine (in this case there is none).