Mandarin Chinese verbs are composed of one to four characters that function together as a conceptual unit. However, in modern spoken Chinese, most of these verbs are two-character compounds. This dissertation is focused on adult second language (L2) learners' acquisition of these two-character verb compounds. Specifically, the acquisition of a morpho-syntactic feature relevant to these verbs: separation. Separation for Chinese verb compounds involves 11 separation patterns in which the two-character verbs are separated and expanded into phrases. For example, in "I swam for one hour," the phrase "one hour" needs to be inserted between the two characters for the verb meaning "swim." This dissertation presents two internet-based experiments. In Experiment 1: 28 learners' knowledge of verb decomposition, grammaticality, and oral translation were tested. Individual difference measures for each participant were also recorded including language background history, estimated lexical knowledge, and estimated grammar knowledge. Results showed that Chinese verb separation is extremely challenging for L2 learners to acquire--even for advanced learners with large vocabularies and extensive grammatical knowledge. When shown a two-character verb, an average of only 37% of the learners could accurately identify the meaning and both characters, i.e., decompose the verb. An average of 63% could identify whether the verbs were used in a grammatical manner or not. In the oral translation task, even if learners were provided with the meaning of the verb compound and the target separation pattern, only an average of 19% could orally produce the correct separation pattern. Namely, a verb was correctly separated in a produced sentence regardless of the grammaticality of the sentence. In Experiment 2: 64 different learners from the same population were trained for 45 minutes using an online language tutor specifically designed to teach verb separation. The tutor included audio-visual input, explicit instruction, visual annotation of prosody, and item-based personalized feedback. Learners were tested with the same three tasks as used in Experiment 1 at roughly two days and again 17 days after the training. The tutor did not lead to an improvement in learners' ability to decompose two-character verbs. However, grammaticality judgment accuracy of the sentences with verb separation improved by 4% in the two-day delayed posttest, and maintained in the 17-day delayed posttest. The mean accuracy of the production of the sentences with verb separation increased drastically by 23% after the training with the tutor. As with the grammaticality judgment task, the learning effect was still present at the 17-day delayed posttest. The dissertation concludes by discussing how the findings contribute to the Unified Competition Model. Additionally, the dissertation discusses its contributions to new experimental methods, i.e., online experiments and L2 pedagogy. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://bibliotheek.ehb.be:2222/en-US/products/dissertations/individuals.shtml.]