“…This is clearly reflected in the frequent use of markedness constraints against such vowels in formal OT (and earlier rule-based) analyses of individual tongue root harmony languages (e.g., Akinlabi 1997, Archangeli & Pulleyblank 2002, Bakovic & Wilson 2000, Cahill 2007, Calabrese 1995, Casali 2003, Leitch 1996, Orie 2003, Pulleyblank 1996, Pulleyblank & Turkel 1996, Wayment 2009). While the degree to which any of these three classes might actually behave as marked in a given language is, under standard assumptions, a consequence of how highly the relevant markedness constraint is ranked, most existing frameworks in which tongue root markedness relations are expressed by means of constraints *[+high,-ATR], *[+low,+ATR] and *[-high, -low,+ATR] or their equivalents provide no straightforward mechanisms under which the expected markedness relations in high, mid or low vowels might be reversed.…”