Morphemes often provide valuable information with respect to the pronunciation of a word. They can also assist linguists with syllable division.
For example, the two English words "finger" and "singer" have different word-initial segments, but this is not the sole difference. The word "singer" consists of two morphemes, the verb "sing" and the agentive suffix -er. The verb "sing" ends with a velar nasal which is also present in "singer". This word consists of five segments or four if one analyzes the -er suffix as a syllabic r or schwar (r-coloured vowel) in which case it has only four. Since no English syllable can begin with a velar nasal, it is clear that singer can be divided as sing + er. The word, finger, however, is different. This word consists of a single morpheme. The "n" is a velar nasal but it is followed by a voiced velar plosive. The word finger consists of six segments or five if one analyzes -er as a syllabic r or schwar (r-coloured vowel). It appears that the only reason "singer" does not have a velar nasal followed by a velar plosive is due to the lack of a velar plosive in the bound morpheme "sing".
In the Spanish word "monstruo" meaning monster, we see a sequence of four consonants in word-medial position. The "u" is pronouned as a "w", a labiovelar glide. Such a consonant sequence does not occur word-initially nor word-finally in Spanish words. Since the consonant sequence -ns does not occur word-finally and -str does not occur word-initially, it is unclear how to divide the two syllables. However, the -ns sequence does occur in proper nouns such as the last name "Sanz" (pronounced with a voiceless dental fricative or voiceless interdental fricative in Castilian Spanish), thus providing evidence for the syllable division mons + truo. A number of Spanish words begin with the consonant sequence tr- such as "tren" (train), "tratamiento" (treatment) and "trono" (throne).
When it is unclear how to syllabify a particular word, evidence from other words can be very useful. In the case of pronunciation, a language may favour a particular pronunciation to remain faithful to the pronunciation of a particular morpheme prior to affixation.