St Mary's was a grand aisleless cruciform church in the 12thc. Much of the surviving building dates from the 14thc and 15thc, including the present aisled nave and both transepts, but the lower stage of the central tower is Romanesque, as are the four arches of the crossing, complete with their carved capitals and supports. Above the W crossing arch, on the W face of the central tower - which is also the internal E wall of the nave - there are two decorated Romanesque round-headed openings, one above the other, on slightly different axes. They light the two-storeyed wall passage running around the tower, and also look down into the nave. The openings now serve the bell chamber above the crossing.

The unaisled Romanesque chancel had been replaced by 1405 by an aisled structure, which was itself demolished in 1541. The massive timber Rood Screen of c. 1480s extends across the full width of the nave and aisles. Important wall-painting associated with the screen partly survives in the nave, above the W arch of the tower.

The only Romanesque sculpture at St Mary’s is found on the capitals of the crossing, the W crossing arch, the bell openings and the interior of the tower.

Attleborough, in the Hundred of Shropham, was held by Thorold before the Norman Conquest and by Roger, son of Rainard, at the time of the Domesday survey. By 1405, the E arm of the church had been demolished and rebuilt, accommodating the college of priests established by the Mortimer family to serve a chantry chapel dedicated to the Holy Cross. In 1541, following the abolition of the English monasteries, this quasi-monastic arrangement led to the demolition of the chancel which, it was claimed, had been appropriated by the college.

Triple shafts on crossing piers also occur at Norwich Cathedral (1096 - 1118), as well as at the Norfolk priories of Binham and Carrow.

The tower openings at Attleborough are sometimes referred to as windows in the literature (eg Pevsner, Norfolk 2, 2000, 185). As they may have been internal features, have no splay and were never glazed, the term 'openings' has instead been used here. The Romanesque openings into the wall passage in the upper stage of the crossing tower at St Mary's, Bampton (Oxfordshire) appear always to have been internal. The Attleborough openings were blocked for a period, presumably in connection with the surrounding wall painting in the 15thc. Pevsner notes that they were reopened in the 18thc.

It is thought that the bells at Attleborough were originally rung from the the ground floor of the crossing and that the present bell chamber is a post-medieval contrivance. It has been suggested that the upper of the two tower openings is a secondary feature, created in the 19c to give the now elevated bell-ringers a good view of proceedings in the nave. The argument seems to depend on the estimated height of the Norman nave roof, and the assumption that it would not have enclosed the upper opening, which could therefore be a later modification. There is no appreciable difference in appearance between the two openings, however. Both seem to have been damaged and rather crudely restored, probably in the course of being blocked and subsequently reopened in connection with the 15thc wall painting.The two carved motifs on the arches of the openings were also used on the W crossing arch immediately below them, namely the double-cone ornament with scalloped clasps, and syncopated double disc (or radial billet), both of which also occur on a regionally distinctive group of doorways surviving at parish churches in the SE of the county (eg Hales, N doorway, Hellington and Heckingham)

Malcolm Thurlby convincingly compares both the stylised foliate cushion capitals and tower wall passages at Attleborough with those in the eastern arm of Norwich Cathedral (Thurlby 1996, 153-4). To these can be added the triple half-shafts on the crossing piers at both buildings. Thurlby suggests a date of c1140 for the crossing at Attleborough. The points of comparison with elements occurring in the first building phase at the cathedral (1096-1118) might suggest a somewhat earlier date for Attleborough.