In advance of articulation, to what extent must the utterance be phonologically planned? Extant evidence suggests that planning is fully incremental (one phonological word at a time), while some studies propose larger units. One of the key reasons is that previous experiments using picture-word interference paradigm have yielded very different results concerning the effects of distractors that are phonologically related to a word in noninitial position of a multiword target utterance. The results reported range from facilitation effect to no effect, and even to inhibition effect. To account for this phonological inhibition effect, Jescheniak et al. (2003) proposed a graded activation hypothesis, which needs further tests. In this study the picture-word interference paradigm was used to test the graded activation account, and to explore the phonological planning scope. Thirty-six undergraduate students participated in the study and were asked to produce utterances of the form “The airplane and the necklace are above/below/on the left/right side of the toothbrush” for objects containing these referents (e.g., an airplane, a necklace & a toothbrush), and as accurately and soon as possible. Each display was accompanied by a auditory distractor which was phonologically related or unrelated to the nouns, or a white noise. The distractor was presents at the same time, or 150ms or 300ms later than the display. All the nouns encompassed two syllables. In the phonologically related conditions for each noun (e.g., “airplane”), the distractor was identical to either the first (e.g., “air”) or the second (e.g., “plane”) syllable. In such a long utterance and elaborate step (i.e., a syllable), the graded activation account was easily explored. In other words, the phonological inhibition effect would be hardly missed if the graded activation account is the truth. The results of speech onset latencies showed a phonological facilitation effect on the first noun of both syllables. In addition, there was no phonological effect on the second and third nouns. There results were contradictory with the graded activation account, and indicated that phonological inhibition effect is not necessary to test the phonological planning scope. The phonological planning scope is the first phonological word, which is consistent with the fully incremental hypothesis. Meanwhile, the results of speech error rates showed a phonological facilitation effect on each noun of the utterances. These results indicated that the phonologically related distractor indeed strengthened the phonological activation of the corresponding noun, and made it easy to be pronounced. However, this facilitation effect was not reflected on the onset latencies for the second and third nouns, which further indicated that the latter two nouns were not phonologically planned before the sentence articulation.